WO2024260575A1

WO2024260575A1 - Detection method and vision device and powder supply system for casting machine

Info

Publication number: WO2024260575A1
Application number: PCT/EP2024/000035
Authority: WO
Inventors: Sabrina Strolego; Stefano Spagnul; Luciano Olivo
Original assignee: Ergolines Lab Srl
Current assignee: Ergolines Lab Srl
Priority date: 2023-06-21
Filing date: 2024-06-18
Publication date: 2024-12-26
Anticipated expiration: 2025-12-21
Also published as: EP4622757A1

Abstract

A vision device for a casting machine comprising an optic device and an acquisition device having a camera, the optic device and the acquisition device being reciprocally arranged one after the other such that light from the optic device is received by the camera, the vision device comprising a fixing system for installation on the casting machine such that, when the vision device is in the installed condition, an opening of the optical device is directed towards a surface of the metal layer in the molten state with a covering of casting powder within the casting machine in such a way that light from the casting machine enters through the opening and is guided to the acquisition device.

Description

DESCRIPTION

DETECTION METHOD AND VISION DEVICE AND POWDER SUPPLY SYSTEM FOR CASTING MACHINE

Technical Field

The present invention relates to a vision device for a casting machine according to the features of claim 1 and a method for detecting the position of the casting powder level.

Prior art

In the field of the production of steel or, in general, of metals and metal alloys, a fundamental role is played by continuous or non-continuous casting machines. In the casting machine, the production process takes place to produce steel semi-finished products called billets, blooms, slabs depending on their size and shape. The production of the semi-finished products takes place from metal or metal alloy in the molten state, which are cast into a mould cooled by means of a cooling fluid flowing in a countercurrent direction to the direction of advancement of the metal semi-finished product, which is gradually formed within the volume of the mould. The mould can be arranged in a vertical or semi-horizontal arrangement. The mould is open at its lower end, from which the semifinished product being formed exits. The mould is open at its upper end, from which the liquid metal gradually begins to solidify within the mould, and then is extracted from the lower end of the mould. The process is stationary, meaning that in the unit of time, an amount of at least partially solidified metal leaves the bottom of the mould which corresponds to the amount of liquid metal entering the top of the mould. Once the casting process has started in the casting machine, the level of the liquid metal within the mould must be kept constant at all times, i.e. the position of the free surface of the liquid metal, i.e. the position of the so-called meniscus, in relation to the inner wall of the mould must be kept constant over time during the process. In order to keep the level of the liquid metal constant, i.e. to keep the position of the meniscus constant, it is possible to act by increasing or reducing the extraction speed of the semi-finished material being formed within the mould, or it is possible to act by increasing or reducing the flow of liquid metal entering the mould from its upper end.

To obtain a measure of the position of the meniscus, so as to control the extraction speed, or so as to control the flow of liquid metal entering the mould, there are commercially several types of sensors today, radioactive, electromagnetic, ultrasonic.

Further, by means of a combination of such sensors, it is also possible to measure and control the thickness of mould powder that is added above the level of the meniscus in order to cover the bath of molten metal or metal alloy within the mould and prevent oxidation processes.

The patent application EP2560774 describes a device and method for controlling the casting powder supply of a continuous casting plant, comprising a first measuring device for determining the height of the bath level in an mould and for generating a corresponding first signal, a second measuring device for determining the temperature of the casting powder on the surface of the bath level and for generating a corresponding second signal, and a computer unit for evaluating the first signal of the first measuring device and the second signal of the second measuring device, by means of which a casting powder supply device can be controlled.

The patent application US6091444 describes a method and apparatus for continuously monitoring a surface of material in a molten state using a camera comprising a CCD detector array, an electronic shutter and an associated optic assembly mounted within a water-cooled enclosure, which is mounted near the surface of material in a molten state. A system of thermal radiation shields surrounds the casing to attenuate the heat radiated by the material in its molten state. To minimize damage to the camera's electrooptic components from vapors and other contaminants, the camera casing includes a small viewing pinhole through which an inert gas is directed. The pinhole is small enough to minimize the flow rate of the gas while avoiding significant image diffraction. The small pinhole also provides a large depth of field, providing high quality images of the surface of molten material. A compound lens system focuses the image onto the CCD detector array to produce a diffraction-limited image relative to angles of interest. The lens system contains a 90-degree rotating element to protect the CCD detector array from direct exposure to X-rays produced on the material surface in the molten state. The electronic shutter, in combination with the pinhole, is used to reduce the incident intensity and to prevent saturation of the CCD during normal use. The shutter speed can be reduced sufficiently to allow alignment with the ambient light. A neutral density filter can also be used in combination with the electronic shutter to prevent saturation.

Patent application EP2363716 describes a device for analysing and determining the movement characteristics of products moving in a certain direction of advancement and emitting radiation, in particular products exiting a casting line, wherein the device comprises a camera for continuously capturing images of the product moving in the direction of advancement in at least two successive instants of time and a processing electronics unit by means of which the comparison is made between at least two successive acquired images, the comparison being made by means of algorithms based on the image correlation principle, in order to determine the spatial displacement of the images and thus the movement characteristics of the product.

Patent application JPS5772752A describes a system for evaluating the need for stopping the supply of molten steel from a ladle to a casting machine by collecting images of the molten steel with a camera equipped with an optic system and detecting the possible inclusion of slag in the molten steel based on the difference in brightness between the molten metal and the slag. Images of the molten steel and the slag floating on it in a ladle are taken with a camera through a lens and an optic fibre, and the images are converted into signals that are displayed on a monitor and sent to an image processing device. A comparator emits an end-of-loading signal when the slag area ratio signal becomes greater than a threshold signal set according to the type of steel and product quality requirements, thereby triggering an alarm and generating a shut-off signal for a flow control valve, preventing excessive amounts of slag from flowing to the mould.

Patent application JP 2002 137049 describes a system for monitoring the level of molten steel in a continuous casting machine based on image processing that is capable of assessing whether the dispersion condition of the casting powder on the molten steel level is good or not. The system monitoring a continuous casting machine molten steel level monitors and controls the level of molten steel being cast into a mould of the continuous casting machine. The level of molten steel on which the casting powder is spread is captured by a camera. The image processor stores the image of the molten steel level when the powder is in good dispersion condition as a reference image, compares the image of the level captured by the camera with the reference image and informs that the dispersion condition of the powder is not good, when a difference between these images is found.

Patent application JP S60 221160 describes a system for enabling rapid detection of the outflow of slag from a ladle into a casting machine by directly monitoring the diameter of the flow of molten metal by optic means and detecting the fluctuation of the diameter of the flow with the molten metal and slag. A sealing element is provided between a ladle snorkel and a tundish inlet pipe in which the wall of the sealing element comprises a side hole provided with a pipe at the end of which a camera is arranged to acquire an image signal of the molten steel flow by means of a signal elaboration unit. The time when the diameter of the molten steel flow abruptly increases is identified as the time when the slag escapes. The elaboration unit applies a closure command to prevent the slag from entering the tundish.

Patent application JP S53 129126° describes a method and apparatus comprising a camera for optically detecting the level of molten metal in a mould of a continuous casting machine. The level sensing signal is used to control the amount of molten metal flowing into the nozzle collar of the tundish or to control the extraction rate of the bar exiting the mould, in order to maintain the level of molten metal in the mould at a constant level. The measuring principle is based on the detection of the difference in brightness between the surface of the molten metal and the surface of the mould wall. The apparatus is equipped with an automatic powder supply system that can continuously supply powder by detecting the brightness of the molten metal surface using the automatic powder supply system. By adjusting the powder supply, the brightness of the molten metal surface is kept always constant to avoid incorrect level measure by the camera due to the variation of the surface brightness.

Problems of prior art

Although there are systems to measure and adjust the production process of metal or metal alloy in the molten state, there is a need to see what is actually happening in the metal or metal alloy bath in the molten state. To date, there is no known device for viewing the metal or metal alloy bath in the molten state in case billets and blooms are used.

The limited available spaces and the extreme environment make it critical to use a camera directly exposed to the irradiation of the molten metal or metal alloy bath. It is also not possible to frame the molten metal or alloy bath using a zoom lens because, for example, with reference to a mould, the presence of the tundish and the casting tray to which the plunger or snorkel is attached prevents the framing of the molten metal or alloy bath which is in a position within the mould, at some distance from the surface of the casting machine with the plunger or snorkel occupying the space available for framing.

As a consequence of these aspects, it is the casting operator who is forced to approach and bend over the bath of metal or metal alloy in its molten state in order to make a direct observation. In some cases, the view is so reduced and the presence of casting equipment is so obtrusive that the operator cannot physically see anything.

Aim of the invention

The purpose of the present invention is to provide a vision device for a casting machine which allows the vision of the bath of metal or metal alloy in the molten state in order to be able to carry out a direct observation even in cases of difficulty of access and particularly reduced spaces.

A further scope of the present invention is to provide a method for detecting the position of the casting powder level.

Concept of the invention

The purpose is achieved with the features of the main claim. The sub-claims represent advantageous solutions.

Advantageous effects of the invention

The solution in accordance with the present invention, through the considerable creative input whose effect constitutes an immediate and not irrelevant technical advance, has several advantages.

The casting machine vision device in accordance with the present invention is an unobtrusive tool which is easy to install without requiring the occupation of space on the casting table and without hindering the ordinary operations of the operators of the casting machine.

The casting machine vision device according to the present invention also allows the camera integrated in the device to be maintained in a protected condition and not exposed to direct radiation from the bath of molten metal or metal alloy.

Definitions

In this description and the appended claims, the following terms are to be understood according to the definitions given below.

The terms "upper", "superiorly", "lower", "inferiorly" are to be understood as referring to the direction of gravity.

In the present invention, the term "liquid metal" is intended to include both pure metals and metal alloys in a liquid state which are at a temperature at least equal to their melting point.

Description of drawings

An embodiment is described below with reference to the accompanying drawings to be considered as a non-limiting example of the present invention in which:

Fig. 1 depicts a perspective view of the casting machine vision device according to the present invention.

Fig. 2 depicts a perspective view of the vision device for casting machine according to the present invention in which internal components are shown in transparency.

Fig. 3 depicts in schematic form a first possible embodiment of the casting machine vision device according to the present invention.

Fig. 4 schematically depicts a second possible embodiment of the casting machine vision device according to the present invention.

Fig. 5 depicts in schematic form a third possible embodiment of the vision device for casting machine according to the present invention.

Fig. 6 is an enlarged view of the portion indicated with A in Fig. 5.

Fig. 7 is a sectional view of a portion of the casting machine vision device according to the present invention according to a further embodiment.

Fig. 8 depicts in schematic form the operating principle of the vision device for casting machine according to the present invention.

Fig. 9 schematically depicts the operating principle of the vision device for casting machine according to the present invention.

Fig. 10 schematically depicts the principle of operation of the vision device for casting machine according to the present invention in which a lighting system is also integrated.

Fig. 11 represents in schematic form the operating principle of the vision device for casting machine according to the present invention in which a lighting system is also integrated according to a different embodiment.

Fig. 12 represents in schematic form the operating principle of the vision device for casting machine according to the present invention in which a lighting system is also integrated according to a different embodiment.

Fig. 13 represents a detail of an embodiment of the vision device for casting machine according to the present invention.

Fig. 14 represents an embodiment of the vision device for casting machine according to the present invention illustrating the light deviation system.

Fig. 15 represents a detail of the vision device for casting machine according to the present invention illustrating a possible cleaning and cooling gaseous flow.

Fig. 16 depicts in schematic form an optic path of the vision device for casting machine according to the present invention.

Fig. 17 represents an embodiment of the vision device for casting machine according to the present invention illustrating a possible gaseous flow for cleaning and cooling.

Fig. 18 represents in schematic form the assembly of the vision device for casting machine according to the present invention.

Fig. 19 depicts in schematic form the assembly of the vision device for casting machine according to the present invention and its integration into a mould powder supply system.

Fig. 20 illustrates the intensity detected by the vision device according to the present invention with reference to a camera acquisition line integrated into the device.

Fig. 21 illustrates one of the processing steps applied on the signals detected by the vision device according to the present invention.

Fig. 22, Fig. 23, Fig. 24, Fig. 25 schematically illustrate the acquisition and processing steps performed via the vision device according to the present invention.

Fig. 26, Fig. 27 illustrate the application of the triangulation principle for the vision device according to the present invention.

Fig. 28 represents an example of acquisition and measure for different installation angles of the vision device according to the present invention.

Fig. 29 represents a further embodiment of the inventive vision device according to a plan view.

Fig. 30 represents a side view of the embodiment of the vision device of Fig. 29.

Fig. 31 represents a front view of the embodiment of the vision device of Fig. 29.

Fig. 32 represents a cooling system usable in the Inventive vision device.

Fig. 33 schematically represents a detail of the optic system side view of the embodiment of the vision device of Fig. 29.

Fig. 34 schematically represents the operation of the second embodiment of Fig. 29, in which the vision device is depicted not to scale and the projected lines are rotated 90° from how they are actually projected with the vision device of Fig. 29, for the purpose of illustrating the principle of operation.

Fig. 35, Fig. 36, Fig. 37, Fig. 38, Fig. 39 illustrate alternative embodiments of one of the components of the second embodiment in Fig. 29.

Fig. 40, Fig. 41 , Fig. 42, Fig. 43 schematically illustrate the image elaboration steps relative to the second embodiment of Fig. 29.

Description of the Invention

The present invention relates to (Fig. 18, Fig. 19, Fig. 29) a vision device (1) for a casting machine (30) and a method of detecting the position of the casting powder level within a mould. The casting machine (30) may be suitable for the production of billets, blooms, slabs, generically referred to as metal semi-finished product (28), and comprises a mould (26) within which a metal or metal alloy in the molten state is cast through a snorkel or plunger (27) that is connected to a distribution tundish of the metal or metal alloy in the molten state contained in a ladle. The mould (26) is cooled by means of a cooling fluid counterflowing with respect to the forward direction of the semi-finished metal product (28), which is gradually formed within the volume of the mould (26). The mould (26) can be arranged in a vertical or semi-horizontal arrangement. The mould (26) is open at its lower end, from which the semi-finished product (28) being formed exits. The mould (26) is open at its upper end, from which the liquid metal enters, which gradually begins to solidify and is then extracted from the lower end of the mould (26) and guided by means of a guiding system (29) combined with a secondary cooling system. The process is stationary meaning that in the time unit, an amount of at least partially solidified metal exits inferiorly from the mould (26) that corresponds to the amount of liquid metal that enters superiorly into the mould (26) itself. Once the casting process is started in the casting machine (30), the level of the liquid metal within the mould (26) must be kept always constant, i.e. (Fig. 18, Fig. 19) the position of the free surface of the liquid metal, i.e. the position of the so-called meniscus (31), with respect to the inner wall of the mould (26) must be kept constant over time during the process. In order to keep the position of the meniscus (31) constant, i.e., to keep the level of the liquid metal in the mould constant, it is possible to act by increasing or reducing the extraction rate of the semi-finished product (28) being formed within the mould, or it is possible to act by increasing or reducing the flow of liquid metal entering the ingot mould (26) from its upper end. The mould (26) is generally oscillating along its axis with reciprocating motion according to (Fig. 18, Fig. 19) an oscillation direction (53).

As previously explained, the limited available spaces and the particularly extreme environment make direct or indirect viewing of the metal or metal alloy bath difficult and critical, i.e., viewing the meniscus (31 ), due to the presence of the tundish and the casting tray to which the plunger or snorkel is attached (27), which occupy the space available for a framing. The use of optic techniques for measure is a good solution. However the difficulties of application requires overcoming the difficulties inherent in the type of process. The main problem is the irregularity of the powder surface, which makes it impossible to use individual measuring points. In fact, the presence of cracks in the powder surface makes the reflection of light extremely uneven. The surface of the powder also has areas with valleys or bulges locally changing the angle of reflection.

In particular (Fig. 1 , Fig. 2, Fig. 18, Fig. 19, Fig. 29), the vision device (1 ) for casting machine (30) is suitable for installation, for example, on the casting table to view the surface of the liquid bath, i.e., the meniscus (31), although other applications at other points of the casting machine (30) or of the production plant are not excluded, for example, for viewing the inside of the tundish or ladle or of a melting furnace. The mounting of the vision device (1) is done by means of a usual fixing system (4), the representation of which is for illustrative purposes only.

The vision device (1) comprises (Fig. 1 , Fig. 2, Fig. 29) an acquisition device (2) and an optic device (3) mutually connected by a clamping system (8) such that light can pass from outside the vision device (1 ) through an opening (10) and enter within the optic device (3) with a first axis (W) of vision, the optic device (3) being configured for light deviation with variation of the first axis (W) of vision and transmission of light to the acquisition device (2) according to a second axis (X) of transmission so that light arrives at a camera (6) of the acquisition device (2). The vision device (1 ) preferably also includes (Fig. 2, Fig. 18, Fig. 19) an elaboration unit (40) for managing the components of the vision device (1) such as the camera (6) and any light emitters (23) or light sources (50), for diagnostics and alarm signal generation, for possible processing of signals acquired by the camera (6), and for managing communication with external devices, such as a central casting machine management unit.

The optic device (3) comprises (Fig. 2, Fig. 29) a case (9) attached to the container (5) by means of a known type of attachment system (8). Within the case (9) are housed:

- a light deviation system (11 ) for light deviation with variation of the first axis (W) and transmission of light according to the second axis (X) of transmission;

- a light transmission system (12) for guiding light along the second axis (X) of transmission to the acquisition device (2).

The acquisition device (2) includes (Fig. 2, Fig. 29) a container (5) within which are housed at least:

- a focus setting system (7) for focusing images of the light received by the optic device (3) in such a way as to focus the image at a point of focus arranged along the second axis (X) on the side opposite to the side on which the optic device (3) is located;

- a camera (6) arranged in such a way that an acquisition sensor is arranged at the focusing point, for image acquisition of light received by the optic device (3) and focalized by the focus setting system (7).

Preferably, the optic device (3) and the focus setting system (7) are connected to the camera (6) in a stable manner creating a single assembly in order to ensure the stability of the overall optic system.

The camera (6) can be a traditional type camera for capturing light in the visible spectrum or possibly it can also be an infrared field sensitive type of camera.

The light deviation system (11) and the light transmitting system (12) are preferably housed (Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 29) within a respective shell (21 ) which has smaller cross-sectional dimensions than the cross-sectional dimensions of the case (9) such that (Fig. 2) within the case (9) there is a free space (13). In practice, the case (9) has, therefore, two functions, one related to the mechanical protection of deviation system (11 ) and transmission system (12) within their respective shells (21) and one related to maintaining the temperature of the viewing device (1) within a range suitable for the operation of the optic and electronic elements, as well as maintaining cleanliness. Within the free space (13) a gaseous flow (44) is supplied (Fig. 17) entering from an inlet (45) arranged on the side opposite to the side of the vision device (1) on which the opening (10) is located, the gaseous flow (44) passing through the vision device (1) and, in particular passing through the optic device (3). The gaseous flow (44) penetrates within the free space (13) and exits through the opening (10), helping to keep a protective glass of the optic device (3) clean as well and keeping the temperature at acceptable values for the operation of the camera (6). Although the optic device (3) and its lenses can operate at high temperatures, up to values of 400°C, the camera (6) should stay within 60°C. For this purpose, the camera (6) preferably includes a first heat sink (43') which is attached to the camera (6) in such a way that heat from the camera (6) is transferred to the first heat sink (43'). Preferably, the first heat sink (43') internally includes a flow channel which is connected to the inlet (45) of the gaseous flow (44). Thus, in general, the vision device (1) may comprise a circuit for a gaseous flow (44) for cleaning and cooling in which the circuit comprises an inlet (45) of the gaseous flow (44), a first heat sink (43') for cooling the camera (6), and a free space (13) for cooling the optic device (3), the gaseous flow (44) exiting the vision device (1) through the opening (10) of light passage. The gaseous flow (44) may be an air flow or an inert gaseous flow such as argon or similar.

For distributing the gaseous flow (44) from the acquisition device (2) to the optic device (3), the vision device (1) includes (Fig. 15) a distribution system (54, 55, 58) arranged at a coupling interface between the acquisition device (2) and the optic device (3). The distribution system (54, 55, 58) includes an insulation flange (58) between the acquisition device (2) and the optic device (3) and a collar (56) spaced apart from the flange (58) forming a chamber (54) for distribution of the gaseous flow (44). The gaseous flow (44) from the acquisition device (2) penetrates within the chamber (54) by means of passages (59) for inputting the gaseous flow (44) within the chamber (54). The collar (56) has radial holes (57) for distribution of the gaseous flow (44) for letting the gaseous flow (44) from the chamber (54) enter within the free space (13) of the optic device (3). Sealing between the flange (58) and the walls of the chamber (54) is achieved by O-rings inserted in corresponding O-ring seats (55).

The entry of light into the vision device (1), i.e., the opening (10) of light entry, cannot be parallel to the optic axis, i.e., the second axis (X) defined by lenses of the transmission system (12), as can be understood from the installation diagram (Fig. 18, Fig. 19) because the available space is small and the vision device (1) cannot be oriented toward the free surface of the liquid metal, i.e., toward the position of the meniscus (31). In addition, it must have a small diameter to decrease the probability of glowing balls of metal hitting the opening (10) of light entry. For these reasons, the optic device (3) includes a light deviation system (11) (Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 8, Fig. 29) which implements the deviation of light by varying the first axis (W) of vision and transmission of light to the acquisition device (2) according to a second axis (X) of transmission arranged horizontally so that the acquisition device (2) and the camera (6) contained therein can be positioned away from the free surface of the liquid metal and protected from direct radiation of heat and splashes of molten metal.

The light deviation system (11 ) (Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 8, Fig. 14) comprises, in the preferred embodiment of the present invention, a pinhole optics (14) coupled with a light-deviation mirror or prism (15), thereby achieving the desired effect of deviation light from the first axis (W) of vision to the second axis (X) of light transmission parallel to the longitudinal extension of the light transmission system (12) to the acquisition device (2) and camera (6). With the pinhole optics (14) coupled with the light-deviation mirror or prism (15), a vision field (22) is obtained (Fig. 9) that widens in a direction toward the surface of the liquid bath at the meniscus (31) such that an extended portion of the liquid bath surface is framed by the camera (6). The pinhole optics (14) is to be understood not as a diaphragm equipped with a hole for light to pass through, but as an actual lens equipped with objective lenses made of optic-type glass. The pinhole optics (14) is designed in such a way that the diaphragm that determines the lens aperture, that is, the factor usually denoted by the number F/n, can be positioned at an inlet of the pinhole optics lens (14). Thus, if the pinhole optics (14) has a short focal length, e.g., 3.3 mm, and an aperture, e.g., of F/3.5, the aperture hole will have a diameter d=3.3 mm I 3.5 = 0.94 mm i.e. , a very small diameter hence the term "pinhole".

For the purposes of the present invention the pinhole optics (14) will have:

- focal length between 1 mm and 15 mm, preferably between 1.5 mm and 5 mm, the optimum and preferred value being 3.6 mm;

- opening between F/1.4 and F/16 preferably between F/2 and F/8, the optimum and preferred value being equal to F/5.6.

Correspondingly, the aperture of the pinhole optics (14) has a diameter between 0.5 mm and 8 mm, preferably between 0.6 mm and 4 mm.

The pinhole optics (14) realizes a vision field (22) that widens in a direction oriented according to the first axis (W) in which the vision field (22) corresponds to a cone of vision with an angular aperture between 10 degrees and 150 degrees, preferably between 60 degrees and 110 degrees, the typical value being about 85 degrees.

If the protective window of the pinhole optics (14) is at this point placed at the aperture, there is the obvious advantage given by a very small window, which is more easily protected from the harsh environment of the mould and exposed to less thermal radiation from the mould (26) and the snorkel (27).

The light transmission system (12) along the second axis (X) can be made in several ways.

In a first embodiment of the light transmission system (12) (Fig. 3), a group of spherical (16) achromatic lenses arranged in couples by obtaining achromatic doublets is used. In this way, it is possible (Fig. 16) to guide the beam of light along the transmission system (12) by means of a series of lens couples (16", 16", 16"") constituting achromatic doublets, such as (Fig. 3, Fig. 16) a first lens couple (16"), a second lens couple (16"), a third lens couple (16"), a fourth lens couple (16""). The example shown (Fig. 16) is for an application in which the light transmitting system (12) has an approximate total axial length of 443 mm, it being obvious to a person skilled in the field that different lengths can possibly be used with more or fewer couples of lenses (16", 16", 16""). In addition, as previously explained, it is envisioned (Fig. 16) the use of the deviation system (11) to deviate light by varying the first axis (W) and transmission of light according to the second axis (X) of transmission. The deviation system (11) may include the previously described pinhole optics (14) combined with a mirror or prism (15). In the case of the light transmission system (12) with a diameter larger than 12 mm, the use of the series of lens couples (16', 16", 16", 16"") constituting achromatic doublets is sufficient for obtaining an image quality suitable for the purpose of the vision device (1 ). Generally speaking, for the application (Fig. 18, Fig. 19) in an mould (26) of a casting machine (30), the vision device (1) can be realized in such a way that the light transmitting system (12) has a diameter greater than or equal to 10 mm, preferably 12 mm, without excessive problems of occupying the casting plane and, therefore, the use of the series of lens couples (16', 16", 16", 16"”) constituting achromatic doublets is a preferred embodiment in view also of the exposure to heat. In this context, further, because of the high temperatures involved, it is preferred that the lens couples (16', 16", 16", 16"") constituting achromatic doublets are not mutually cemented lens couples but are mutually fixed lens couples with air spacing, that is, in which air is present between the lenses constituting the lens couple.

For applications in which the transmission system (12) of light must be less than 10-12 mm in diameter, in order to obtain an image quality suitable for the purpose of the vision device (1 ), it is preferred to adopt a transmission system (12) other than the one with the series of lens couples (16', 16", 16", 16"") constituting achromatic doublets previously described, such as:

- transmission system (12) realized (Fig. 4) by means of rod lenses (17) or

Hopkins lenses; - transmission system (12) realized (Fig. 5, Fig. 6) by means of relay lenses (18) or GRIN lenses, i.e., with a refractive index gradient, in which the relay lenses (12) are arranged along the second axis (X) and alternated with air gaps (19).

Alternatively, the transmission system (12) may be realized by the use of coherent bundle of receiving optic fibres. In this case, the use of high-temperature receiving optic fibres made of glass and not of the polymer type is provided, thus being able to obtain a flexible transmission system (12) that allows the vision device (1) to be positioned more easily even in conditions where space is particularly tight. This embodiment, while preferred from the point of view of practicality of use, has a disadvantage with regard to the small number of pixels obtainable by means of the receiving optic fibres that are made available for acquisition by the camera (6). In fact, with a transmission system (12) at the receiving optic fibres a number of pixels of approximately a hundred thousand can be obtained, unlike the other described solutions that allow resolutions in the order of megapixels. In any case, the reduction in image definition is acceptable when maximum installation flexibility is required.

In general, the choice of one embodiment of the transmission system (12) over another among those illustrated depends on the cross-sectional size of the shell (21) and the case (9). As explained, the preferred embodiment of the present invention uses a light transmission system (12) comprising a series of lens couples (16", 16", 16"") constituting achromatic doublets with a diameter greater than or equal to 10 mm, preferably 12 mm. This value also represents a limitation due to the amount of light that can be collected compatible with the dynamics of the process. In fact, since it is necessary to acquire by means of the camera (6) images at a frequency of at least 3 Hz, the exposure time cannot be increased at will but must be limited within 100 ms, better under 50 ms, otherwise the presence of flames on the free surface of the liquid metal, i.e., at the position of the meniscus (31), and the oscillatory movement of the mould (26) according to (Fig. 18, Fig. 19) a direction of oscillation (53) that induces oscillations of the free surface of the liquid metal tend to reduce the quality of the image.

Indicatively, the length of the optic device (3) can be, for example, between 250 mm and 400 mm, preferably about 300 mm, but of course these are exemplary values. The length of the optic device (3) cannot be much less than this range, as the thermal stress induced on the camera (6) would be excessive. The length of the optic device (3) cannot be much greater than this range as installation would be made difficult.

In an optional embodiment, it is provided (Fig. 7, Fig. 10) for the vision device (1) to also include a lighting system (23, 24) comprising a light emitter (23), for example by means of LEDs, wherein the light emitter (23) may be arranged within the container (5) of the acquisition device, and wherein the light emitter (23) optically couples with a light transmitter (24) by transmission of the light generated by the light emitter (23) to the opening (10) in order to illuminate the surface of the liquid bath at the meniscus (31).

In some situations, in fact, the light emitted by the snorkel (27) or incandescent plunger and the ambient light are not sufficient to obtain a sharp image from the camera (6). The introduction of the active illumination function allows this problem to be solved. It will be obvious to an expert in the field that several solutions will be possible to achieve illumination. The main factor is that the light emitter (23) combined with the respective light transmitter (24) be such that uniform illumination of the liquid bath surface at the meniscus (31 ) is achieved to improve the quality of the image captured by the camera (6).

In some embodiments, the lighting system (23, 24) is arranged (Fig. 10) inside the case (9) or inside (Fig. 7) the shell (21). For example, the light transmitter (24) can be made (Fig. 7) by means of optic fibres (20) arranged radially around the light transmission system (12) for guiding light along the second axis (X) of transmission. In this case, the light transmitter (24) can be realized in the form of a cylindrical beam of incoherent optic fibres inserted into the free space (13) between the case (9) and the shell (21).

Since the lighting system (23, 24) is considered as an optional element and also in order not to have problems of internal reflections of light from the lighting system (23, 24) that could affect the acquisition by capturing light through the opening (10), the lighting system (23, 24) is preferably realized as an external system with respect to the case (9) of the vision device (1).

The light emitter (23) is preferably a LED diode coupled to the cylindrical bundle of incoherent optic fibres in which the LED diode is controlled in such a way as to send light pulses synchronized with the camera acquisition time (6). Pulsed illumination advantageously allows the LED diode to be employed at much higher power during the useful exposure time set on the camera (6), as opposed to the case of continuous driving of the LED diode, which would result in unnecessary heating of the LED diode even under conditions where illumination is not needed in relation to the acquisition frequency and exposure parameters of the camera (6). Additionally, pulsed illumination eventually allows the acquisition, elaboration and comparison of successive images alternating with and without illumination, allowing for better discrimination between the surfaces in the image, such as the snorkel surfaces (27), mould (26), powder covering the molten state metal bath present in the mould (26) itself. For this purpose, the camera (6) includes trigger inputs to control the acquisition and strobe control outputs to control the LED diode to turn on in synchrony with the image acquisition. In fact, the snorkel (27) is generally at a temperature of 800/900 °C, and most of the blackbody emission occurs in the red and green field. This applies also for flames developing on the surface of the bath.

This provides (Fig. 10) an illumination field (25) at least partially overlapping the vision field (22) of the vision device (1).

In one embodiment (Fig. 18) it is envisaged that the vision device (1) can be installed independently and autonomously. In other embodiments (Fig. 19), it is envisaged that the vision device (1) can be integrated with a powder supply system (32) for mould (26), which provides for delivering and distributing on the surface of the liquid bath at the meniscus (31 ) casting powder that is intended to keep the liquid bath protected from oxidation and also to act as a lubricant for sliding along the walls of the mould. In fact, in the continuous steel casting process, cover powders play a key role, and the reasons for their use are many and interrelated. First of all, these powders offer valuable thermal protection: they form an insulating layer on the surface of the mould, which helps retain the heat of the liquid steel during solidification. This is crucial to ensure that the steel solidifies evenly and that the resulting mould has a homogeneous internal structure and desired properties. In addition, the cover powders act as a kind of shield against oxidation. They protect the steel from reaction with atmospheric oxygen, preventing the formation of surface oxides that could compromise the quality of the ingot. This is especially important for high-precision steel grades, where surface purity is critical. At the same time, these powders can influence the solidification rate of the steel, helping to control the formation of dendrites and achieve a uniform grain structure of the ingot. This implies that they not only protect the steel but also influence its final microstructure, which is essential to the mechanical properties of the material. Another important aspect is the reduction of friction. Cover powders reduce friction between the liquid steel and the mould walls, thus allowing for smoother and more uniform casting. This helps to prevent problems such as mould adhesion and ensures continuous and uninterrupted production. Finally, cover powders can play a role in controlling unwanted inclusions in steel. Acting as binding agents, they can capture and retain harmful inclusions, helping to improve the purity of molten steel and reducing defects in finished products. Controlling the thickness of cover powders is essential to ensure effective action, and given the difficulty of general direct control by man of the amount of powder supplied into the mould.

Therefore, in this case, a powder supply system (32) with an integrated vision device (1) can be obtained. The vision device (1 ), in this case, includes a conduit (33) to supply a mixture of conveying air and casting powder to one or more casting powder spreader heads (34) to the mould (26). The mixture of conveying air and casting powder is supplied by means of a hose connected to a pneumatic system (36) in which the mixture of powder from a tank (35) and air is formed for subsequent injection of the mixture of powder and pressurized air within the conduit (33). A control unit (37) will be able to communicate via a first communication channel (38) with the elaboration unit (40) of the vision device (1), for example, to receive signals of the absence of casting powder or the presence of an insufficient amount of casting powder or meniscus level position signals (31) for comparison with steel level signals, in the mould to determine a measure of the thickness of the casting powder in the mould for its adjustment by the control unit (37). The control unit (37) will be able to further communicate via a second communication channel (39) to control the supply of powder from the tank (35) by starting the pneumatic system (36). This solution is particularly effective because it allows having two functions by means of a single object installed on the mould (26) which facilitates installation even in case of limited available spaces and further facilitates the preparation set-up of the casting machine (30) since by moving a single object both the vision device (1) and the end part of the powder supply system (32) are installed.

Thus, the present invention also relates (Fig. 19) to a powder supply system (32) for a mould (26) for delivering casting powder within the mould (26), wherein the powder supply system (32) includes a control unit (37), a tank (35) and a conduit (33) for supply of a mixture of conveying air and casting powder to one or more casting powder spreader heads (34) of casting powder to the mould (26) wherein the powder supply system (32) includes a vision device (1) as previously described. To this end, the elaboration unit (40) is configured for transmitting the measure of the position of the beaming line (42) to the control unit (37) via a first communication channel (38) of the communication media, the control unit (37) being configured for receiving the measure of the position of the beaming line (42) and being configured for receiving a measure of the position of the steel level in the mould casting from a level sensor (62), the control unit (37) being further configured for calculating a measure of the thickness of the casting powder in the mould based on measure of the position of the beaming line (42) and measure of the position of the steel level, the control unit (37) being configured for calculating a control signal for generation of a command for supply of powder into the mould from the tank (35). The level sensor (62) may be a level sensor of a known type, such as an electromagnetic, radioactive, or other equivalent type of steel level sensor in the mould.

In both the embodiment (Fig. 18) in which the vision device (1) is independently installed and in the embodiment (Fig. 19) in which the vision device (1) is integrated with the powder supply system (32), it is provided that the attachment on the mould (26) will be done using a magnetic type of attachment system (41) granting extreme flexibility to the casting operator who has to arrange the equipment after the continuous casting machine has started.

With special reference to the processing by the elaboration unit (40) of the signals acquired by the camera (6), it should be noted that, from the analysis of RGB images acquired by an acquisition camera oriented so as to frame the surface of the liquid metal at the meniscus (31), and, in particular, from the analysis of the individual color bands constituting the RGB signal, it is shown that in the band corresponding to the frequency of the color blue the emission of black body of the snorkel (27) or plunger and the emission due to surface flames is particularly small. In the band corresponding to the frequency of the red color, the inner surface of the mould (26) and that of the overlaying powder are clearly seen, but in this context, the presence of surface flames quickly makes the acquired image saturated by obscuring the view of the powder level. In the blue band, on the other hand, with the same exposure, it is basically impossible to determine the powder level due to poor illumination. For this reason, active illumination by means of the lighting system (23, 24) and in particular with reference to the band corresponding to the frequency of the blue color, allows for a contrast similar to that of the red band, but without the disturbance due to saturation caused by the presence of any open flames on the surface of the liquid metal at the meniscus (31).

The vision device (1) has two objectives:

- to identify the position of the powder level (L1) in the mould (26), i.e., the position of the dividing line between the mould and the casting powder covering the liquid metal surface at the meniscus (31);

- to provide an operator with an image for direct supervision of what is happening within the mould (26).

The identification of the position of the powder level (L1) enables the interfacing of the vision device (1) with a steel level meter (L2) within the mould (26), thus enabling a measure of the powder thickness in the mould to be obtained by means of the difference between the powder level (L1) and the steel level (L2), thereby being able to

- generate an alarm signal if the thickness of the powder in the mould is found to be below a certain minimum threshold;

- generate an alarm signal if the thickness of the powder in the mould is found to be above a certain maximum threshold;

- regulate the supply of powder into the mould by the powder supply system (32), whether it is made in an integrated form with the vision device (1 ), or whether it is a separate and autonomous system with respect to the vision device (1).

Additionally, by analysing the image acquired by means of the vision device (1), it is possible to identify the presence of any areas of the surface that are covered by powder and thus leave the metal bath exposed in its molten state, which can be indicative of excessive powder consumption or poor powder supply flow by the powder supply system (32), such situations being able to adversely affect the quality of the final product.

In an embodiment (Fig. 11, Fig. 12, Fig. 13), it is contemplated that the viewing device (1) has a projection system (46, 50, 51) of a beaming line (42) of light which is projected within the vision field (22) of the vision device (1). The beaming line (42) is projected by the vision device (1) and is presented on the surface of the liquid metal at the meniscus (31), i.e., the beaming line (42) is projected by the vision device (1) onto the layer of mould powder covering the underlying molten metal. The elaboration unit (40) is configured for processing the series of images (11 , I2, I3) acquired by means of the camera (6) while obtaining a measure of the position of the beaming line (42) disposed superficially on the cover of molten metal casting powder, i.e., the elaboration unit (40) being configured for calculating a measure of the position of the level (L1) of the powder in the mould. The vision device (1) includes communication means for transmitting the measure of the position of the beaming line (42) and/or the measure of the position of the level (L1 ) of the powder in the mould. The means of communication are of a type known to one skilled in the art, being capable of being realised by means of an analogue type current or voltage output, by means of a data connection, such as a serial type connection, by means of a wired or wireless type network connection.

In a first embodiment (Fig. 11) of the vision device (1) with projection system (46, 50, 51 ), it is envisaged that the projection system (46, 50, 51) is integrated internally into the vision device (1) and the beaming line (42) is projected from the vision device (1) through the light inlet opening (10) for the camera (6).

In a second embodiment (Fig. 12) of the vision device (1) with a projection system (46, 50, 51), it is envisaged that the projection system (46, 50, 51) is integrated externally to the vision device (1 ), such as laterally or superiorly, in which case the beaming line (42) being projected directly from the projection system (46, 50, 51) itself without affecting the light entry opening (10) for the camera (6), with the desired triangulation angle.

For example, the projection system (46, 50, 51 ) comprises a light source (50), such as a laser diode, a light guide (51 ) and a projector (46). Preferably, a laser diode coupled with an optical fibre is used which directly makes a light source (50) and light guide (51) assembly. The optical fibre may be singlemode or multimode, depending on the power and type of laser diode. A related problem is the blackbody emission, which is particularly intense as the temperatures are in a range whose upper limit can be up to around 1500°C. Finally, the presence of flames generates optical emission points on the surface of the powder, which occur continuously and are randomly distributed. To overcome these problems, it is necessary to use a light source (50) in the blue band. Furthermore, in order to overcome the aforementioned problems relating to the unevenness of the surface of the powder in the mould, the light source (50) must not be point-like but must have a predominant dimension, e.g. a stripe, so that it can cover a large area of the surface of the powder in the mould. In this way it is possible to reconstruct missing points due to local imperfections in the surface of the powder.

For example, for a blue laser diode having a wavelength of the order of 450 nm and a power of the order of 15 - 20 mW, the optical fibre will typically be a single-mode optical fibre. For example, for a blue-violet laser diode having a wavelength of the order of 405 nm and higher powers than those indicated above, such as a power of the order of 300 mW, the optical fibre will typically be a multimode optical fibre with a core diameter of 50 micrometres. The projector (46) comprises a collimator (48) which collects the divergent beam of light exiting the optical fibre constituting the light guide (51) and collimates or focuses it at the working distance, a line generating optic (52) and a reflecting surface or prism (47) to correctly orient the projected line towards the mould powder layer at the desired triangulation angle. The line generation optics (52) may be a cylindrical lens, a Powell lens or diffractive optics.

Preferably the beaming line (42) of light is blue, since the frequency of light corresponding to the colour blue is, as explained above, more detectable and not subject to saturation problems in the camera as is the case with light corresponding to the colour red. A blue-coloured line projected onto the powder remains very visible and distinguishable even with infra-red radiation due to the glowing snorkel and any open flames or exposed steel due to the low quantity of powder in the mould. In general, the beaming line (42) is a line of light having a wavelength between 400 nm and 480 nm. As an additional measure to avoid saturation of the camera acquisition sensor (6), both for the first embodiment (Fig. 2, Fig. 3) and for the second embodiment (Fig. 29, Fig. 30), a filter (63) preferably an interference-type filter is in any case provided. In case of using a light source (50) that has, as explained above, a wavelength between 400 nm and 480 nm, the interference filter is a 450 nm filter with a bandwidth of at least 25 nm to allow for tolerances of the wavelength generated by the light source (50), the centre-band of the filter (63) itself, and also the variation of the wavelength generated by the light source (50) due to temperature variations. By using very low tolerances, one could decrease to a filter bandwidth (63) of 10 nm, whereas, by driving the laser-type light source (50) with more current, it is advisable to have a greater filter bandwidth (63), reaching a bandwidth of at least 40 nm. The required out-of-band optical density is OD=4, i.e. the out-of-band attenuation is 10⁴. Modifications and adaptations in relation to possible different wavelengths of use for the light source (50) will be immediately apparent to one skilled in the art, which are however to be considered possible with respect to the shown preferred solution and which are in any case to be considered within the scope of the present invention.

In this case, a region of interest (ROI) is identified (Fig. 11 , Fig. 12) with reference to which the image acquired by the camera (6) is examined to identify the points of the beaming line (42) and then a triangulation principle is used to obtain a measure to identify the position of the beaming line (42), i.e. the position of the powder level (L1) onto which the beaming line (42) is projected.

By framing with the camera (6) the beaming line (42) projected by the vision device (1) on the layer of mould powder covering the underlying metal in the molten state and considering the intensity detected on a row (or column) of a frame acquired in the region of interest (ROI), it can be seen (Fig. 20) that scrolling through the pixels (P) that make up the acquired image, the variation in intensity (I) due to the presence within the field of vision (22) and, in particular in the region of interest (ROI), of the beaming line (42) is clearly evident, thus making it possible to identify the pixels in correspondence with which, for a series of rows or columns of a frame acquired, the beaming line (42) is present. The trend depicted (Fig. 20) is that of the average intensity detected in the row (or column) of a frame acquired in the region of interest (ROI). It will be obvious to a person skilled in the field that many different algorithmic approaches can be used to extract the spatial information regarding the average position of the beaming line (42) within the row (or column) of the acquired frame. In the present description, by way of example, the identification method based on the identification of the value (on the ordinates of Fig. 20) of the maximum intensity (I) is used for simplicity, from which the corresponding position index given by the pixel (on the abscissas of Fig. 20) at which there is the intensity peak is derived. It will be evident, however, that when in the present description we refer of the identification of the spatial position of the beaming line (42) within the acquired frame, various known algorithms of the prior technique can be applied, among which that based on the principle of maximum intensity. Repeating (Fig. 21 ) this operation for a sequence of frames acquired by the camera (6) is equivalent to repeating this operation in time, as each frame acquired corresponds to an acquisition made by the camera at a given instant in time. Consequently, by identifying the spatial position, i.e. the pixel at which the beaming line (42) is located for each frame of a sequence of acquired frames (Fig. 21) and reporting this data in ordinate as a function of the corresponding number of frames (i.e. as a function of time), it can be seen that the pixel in correspondence of which the beaming line (42) is located shifts due to the vertical oscillation of the mould (26), which is carried out to make the extraction Of the product from the mould easier, according to (Fig. 18, Fig. 19) an oscillation direction (53). In fact, while the powder level (L1) is kept essentially in the same position, the mould (26) and the vision device (1) mounted on it oscillate vertically in such a way that the vision device (1) moves away from and approaches the powder level (L1 ) with a known excursion range that can be, for example, of the order of 7-10 mm. By analysing the sequence of acquired frames (Fig. 21), it can be seen that the pixel/mm ratio is roughly of 1 , which is a very good result in order to be able to evaluate powder surface ranges of a couple of tens of millimetres, enough to make an automatic control of the flow of powder introduced by means of the powder supply system (32). In the specific example (Fig. 21), the frame acquisition frequency is approximately 10 Hz and thus it can be deduced that the period of oscillation of the mould (26) is in the range of 2-3 Hz compatible with the typical process value, thus proving the validity of the measure.

The transformation from pixels to mm can be done by using the triangulation method as mentioned above, or simply by means of a conversion table obtained from an appropriate calibration procedure.

Although in the figures and description reference has been made to a beaming line for explanatory purposes and for simplicity, it will be evident that more complex solutions can be used, such as a projection pattern in grid form, according to techniques known to a person skilled in the field, provided that the projection pattern allows for measurable contrast on the frame acquired by the camera (6). It is considered particularly advantageous, for the reasons stated above, that the light beaming line or pattern is projected by means of a projection light beam at a frequency corresponding to that of the colour blue.

With particular reference to the principle of triangulation (Fig. 22, Fig. 23, Fig. 24, Fig. 25) for the precise identification of the position of the powder level (L1) from the acquisition of the light beaming line or pattern (42), the method comprises the following steps:

- acquisition (Fig. 22) of a series of images (11 , I2, I3, In) over time by means of the camera (6) of the vision device (1 ) in which each image comprises (Fig. 23) the beaming line (42) projected by the projection system (46, 50, 51 );

- extraction - preferably, but not necessarily - from each of the acquired images of only the blue colour channel of the RGB composition of the image, this step being advantageous since the frequency of the light corresponding to the blue colour is more detectable and distinguishable in the presence of the infrared radiation of the snorkel (27), mould (26) and molten metal;

- analysis of each of the acquired images by application of a luminous intensity threshold with identification (Fig. 23) in the series of acquired images of a series of points of higher brightness intensity

- possible filtering of each of the acquired images with identification in the series of acquired points of higher brightness intensity of mutually aligned points corresponding to the acquired beaming line (42) and elimination of non-aligned points corresponding to points of higher brightness intensity not related to the beaming line (42);

- for each of the images application of a linear regression function with identification of lines acquired in single projection (A1, A2, A3, An), such as (Fig. 23) a first line acquired in single projection (A1) from a first image (11), a second line acquired in single projection (A2) from a second image (12), a third line acquired in single projection (A3) from a third image (13), an n^th line acquired in single projection (An) from an n^th image (In) wherein each acquired line corresponds to the acquisition of the beaming line (42) at a different position of the vision device (1) with respect to the powder level (L1 ) wherein the different position is caused by the oscillation along the direction of oscillation (53) of the mould (26) and of the vision device (1) attached to the mould (26);

- for each of the lines acquired in single projection (A1, A2, A3, An) calculation (Fig. 25) of a midpoint by obtaining a series of midpoints (P1, P2, P3, Pn) of the lines acquired in single projection (A1 , A2, A3, An), such as a first midpoint (P1) of the first line acquired in single projection (A1) a second midpoint (P2) of the second line acquired in single projection (A2), a third midpoint (P3) of the third line acquired in single projection (A3), a n^th midpoint (Pn) of the n^th line acquired in single projection (An);

- calculation of the position of the level (L1 ) of the powder by application to the series of midpoints (P1, P2, P3, Pn) of trigonometric triangulation formulas (Fig. 24, Fig. 25) or by means of a previously calculated correspondence table, taking into account (Fig. 24) of the distance constituting the basis of triangulation (S) between the vision point (PV) corresponding to the position of the deviation system (11) which sends the light to the camera (6) and the projection point (PP) of the projector (46) with compensation of the distance variations (DP1 , DP2) of the series of midpoints (P1, P2, P3, Pn) due to the oscillation along the direction of oscillation (53) of the mould (26) and the vision device (1) attached to the mould (26).

It will be obvious to a person skilled in the field that other calculation methods can be used by using the general principles of optical triangulation. As an example, for a height (H1 , H2, H3) of the vision device (1 ) relative to the position of the powder level (L1 ) of 150 mm and a measure accuracy of 1 mm, a triangulation base of no less than 20 mm can be assumed.

The distance constituting the triangulation base (S) between the vision point (PV) corresponding to the position of the deviation system (11 ) that sends light to the camera (6) and the projection point (PP) of the projector (46) could be, for example, in the range 20 mm to 50 mm.

Considering (Fig. 26, Fig. 27) the application of the principle of triangulation for the vision device according to the present invention, it is found that the triangulation base (S) is equal to the distance between (Fig. 26) the points BO.

Considering (Fig. 26, Fig. 27) a first height (H1 ) and a second height (H2) of the vision device (1 ) with respect to the position of the powder level, due to the oscillation of the mould, the following relationship between the angles of vision with respect to the position of the beaming line (42) is obtained: (1)

The distance (Fig. 26) between points B and H1 is equal to: (2)

By applying known trigonometric formulae, we have: sin(ψp)= sin(n/2 - (3)

co(ψp) = cos(n/2 (4)

Combining equation (3) with equation (4) in the tangent formula gives: (5)

It can be seen from the figures (Fig. 26) that the relationship holds:

= C/E (6)

Combining equation (6) with equation (5) gives: (7)

Since (Fig. 27) the second height H2 is equal to the distance between the points O and R2, we have:

)) (8)

Being further: (9)

The desired second height H2 is obtained:

H2= S

In equation (10) S denotes the triangulation base as defined in the figures, it indicates the angle of inclination of the device with respect to the casting plane corresponding to the first angle (AG1), C is the beaming line bearing detected by the camera sensor (6), it indicates the angle between the first axis (W) and the axis of the beaming line i.e. the angle between the projection point (PP) and the vision point (PV), E is a known construction parameter of the vision device (1) in relation to its design dimensions such as the distance of the camera from the deviation system (11 ).

If we consider (Fig. 28) an application example with triangulation base (S) equal to 50 mm, the angle between projection point (PP) and vision point (PV) i.e. the angle between first axis (W) and beaming line axis equal to 12.5°, focal length of 3.5 mm, pixel size of camera sensor (6) equal to 3. 45 pm, usable area of the camera sensor (6) equal to 174 pixels x 174 pixels, it can be seen that the height measure (H) corresponding to the position of the beaming line on the mould powder layer expressed in mm can also be obtained on the basis of the curves depicted for different possible installation angles, corresponding to the first angle (AG1). In Figure (Fig. 28) the curves for 0°, 15°, 30° are shown as examples. The typical working area for a mould installation for the powder thickness control application lies approximately between 25 and 50 pixels of the camera sensor (6), corresponding to a height (H) between 150 and 200 mm. In the figure (Fig. 28), the pixel marked 'O' represents the centre of the camera sensor (6).

In a further embodiment (Fig. 29, Fig. 30, Fig. 31 , Fig. 32, Fig. 33, Fig. 34) of the vision device (1), at least two projection patterns (42', 42") are provided, such as a first projection pattern (42') and a second projection pattern (42"). In the illustrated embodiment (Fig. 34), the first projection pattern (42') is made in the form of a first beaming line (42') and the second projection pattern (42") is made in the form of a second beaming line (42"). However, it is contemplated that the present invention is not limited to the illustrated specific case of projecting two beaming lines, as it may be contemplated that the projection patterns (42', 42") are more complex than the illustrative case of beaming lines. Indeed, it may be envisaged that each of the projection patterns (42', 42"), at least two of them, may be selectable from a pattern in the form of a line, a pattern in the form of a circumference, a pattern in the form of two mutually connected segments angled with respect to each other, a series of parallel lines, or other more complex patterns which will be immediately apparent to a person skilled in the field of laser beam generation, e.g. by diffractive optics or the like. What is important for the further embodiment described (Fig. 34) is that at least two projection patterns (42', 42") are present which are not mutually parallel but are mutually convergent, as will be explained later in this description. However, with respect to the listed possibilities the embodiment which envisages a first pattern in the form of a first beaming line (42') and a second pattern in the form of a second beaming line (42") is more advantageous in that it does not require the use of more complex and cumbersome generating optics, losing the essential advantage of being able to have a compact vision device (1 ) suitable for installation in a harsh installation environment such as a continuous casting machine on which there is little space available and compact dimensions are also preferred for protection and safety issues of the measuring instrument itself. As with the previously illustrated embodiments, the projection patterns (42’, 42"), in this case converging patterns, are projected (Fig. 34) within the vision field (22) of the vision device (1). The projection patterns (42', 42") are projected by the vision device (1 ) and are shown on the surface of the liquid metal at the meniscus, i.e. the projection patterns (42’, 42") are projected by the vision device (1) onto the layer of mould powder covering the underlying molten metal. In the specific case of the first pattern in the form of a first beaming line (42') and the second pattern in the form of a second beaming line (42"), two slightly converging laser beams intersect the surface of the powder, generating two beaming lines (42', 42") which can be acquired, as previously explained, by the camera (6) of the vision device (1). Since the two laser beams are slightly convergent, the spacing measure (M1 , M2, M3) between the beaming lines (Fig. 34) resulting from the projection depends on the relative position of the powder surface and the laser source. This spacing measure (M1 , M2, M3) is the greater the closer the powder level (L1) is to the vision device (1 ), i.e., the higher the powder level (L1) in the mould (26). Between the spacing measure (M1 , M2, M3) between the beaming lines (42', 42") and the powder-source distance, there is a precise geometric relationship that allows the latter to be derived as a function of the former, which is a measure that can be obtained from the series of images (11, I2, I3) acquired by the camera (6) of the vision device (1). Advantageously, the projected laser beams have a straight extension allowing the use of linear interpolation techniques for beam reconstruction during the analytical measure process.

The principle behind the measure of this further embodiment (Fig. 29, Fig. 30, Fig. 31 , Fig. 32, Fig. 33, Fig. 34) can be described in terms of perspective effects. First, it should be considered for the mere purpose of exemplifying the principle, a couple of laser beams parallel to each other, with a distance between the two parallel beams that is known, in which the two parallel laser beams hit the plane of a screen perpendicularly resulting in the projection of two mutually distant points of the same distance between the two parallel beams. If the screen is observed from a certain distance, with the observation point placed on the same direction from which the two beams are projected, it will be seen that, as the distance of the observation point from the screen varies, the distance observed on the screen between the two points projected on the screen will also vary proportionally to the variation of the distance of the observation point due to the perspective effect. If the screen is observed by a camera and the digital image acquired by a computer, it is possible to derive the screen distance from the distance between the two projected points. Similarly, consider, however, the case where, instead of two points, two straight line segments are projected onto the screen, paired and parallel, i.e., a first beaming line (42') and a second beaming line (42"), limited to this example illustrative of the principle, the beaming lines (42', 42") being the result of a projection of mutually parallel beams and not converging as previously explained. Similar to what has been illustrated for two-point projection, again from the measure of the distance between the two projected lines, a measure of the distance of the observation point from the screen can be obtained. Although there are, theoretically, no differences from the case of two-point projection, the use of two beaming lines instead of two projection points makes it possible to obtain more precise measures because statistical methods can be applied to reduce errors due to the fact that lines are acquired that allow for multiple acquisition points for each of the projected lines. The determination of powder level (L1 ) in mould (26) requires that distances of less than a meter are typically measured. Under such conditions, it is useful to converge the two laser beams, symmetrically with respect to a common middle and bisector axis, for example coincident with the previously defined first axis (W) of vision device (1 ), according to a configuration in which the two laser beams converge to a single point or rather in a single line in case of two beaming lines, and in which the point of convergence is located a few centimeters beyond the area of greatest distance that is to be measured, that is, a few centimeters beyond the minimum position of the powder level (L1 ) in the mould (26). In this way, i.e., by bringing to a finite distance the point of convergence, which in case of non-converging parallel projection is placed at infinity, the perspective effect is greatly amplified, facilitating and making more precise the measure of the powder level (L1) in the mould (26). For this reason, the further embodiment illustrated (Fig. 29, Fig. 33) includes, in addition to the light source (50), having the same characteristics as previously described, also a beam splitter (65) for splitting the single beam generated by the light source (50) into two mutually parallel beams and, subsequently along the optical path, a component called the perspective amplifier (66) which is suitable to give the described convergence configuration between the two beams. Specifically (Fig. 29, Fig. 33), the light source (50) is preferably a multimode laser diode that is controlled in such a way as to send light pulses synchronized with the camera acquisition period (6) or a camera shutter (6). The light source (50) in the form of a multimode power laser diode has its own "slow-axis" by which we mean the axis of the laser diode having less divergence from the other axis, called the "fast-axis" of the laser diode itself as a result of emitter asymmetry. Preferably, the light source (50) is arranged with its fast-axis parallel to the plan displaying plane (Fig. 33). The preferred power of the light source (50) in the form of a laser diode is between 100 mW and 5 W, preferably about 1 W. Preferably the wavelength of the light source (50) in laser diode form is between 390 nm and 525 nm, preferably between 400 nm and 480 nm, the preferred value being, as previously explained, 450 nm. Other usable and advantageous values for which power laser diodes are currently available are 405 nm for greater contrast than thermal emission wavelengths and 525 nm for greater visibility to the naked eye. However, wavelength values in the visible and nearinfrared range are possible. Alternatively, a single-mode laser diode or a diode-pumped solid-state laser could also be used but without obvious advantages. Other types of lasers may be excluded due to space requirements. The factor relating to the beam pointing stability of the laser is not an important factor as it is in a conventional laser profile meter that makes an absolute measure, since, as explained, in this case a differential measure principle is used, due to the fact that a distance between two projected beams is measured. This choice causes fluctuations in the beam direction to act together on the final measure by canceling out due to differential measure.

Still considering the optical path, downstream (Fig. 29, Fig. 33) of the light source (50) there is the collimator (48), which collects the light beam exiting the light source (50) and collimates or focuses it at the working distance. In this case, the collimator (48) is preferably a collimating lens with a focal length of 6.2 mm, numerical aperture 0.40 and is preferably an aspherical lens. However, in view of the non-criticality of the quality of the beam emitted, it could be replaced by simpler and cheaper optical systems, as immediately apparent to a person skilled in the field. The focusing operated by means of the collimator (48) is preferably such that the beam is focalized in such a way as to emit a "beam waist" within the vision field (22) of the vision device (1), near the proximal end, but the beam could also simply be collimated. The focal length could be shorter, as short as 2 mm, or as long as 8 mm. The advantage of a long focal length as employed is easier focusing and centering than the laser diode. The advantage of a shorter focal length might be less wasted light energy at the next slit (64) but we must consider the relatively large length, in the slow-axis direction, of the emitting area of the power laser diode, which places more stringent constraints on the optical design of the collimating lens for short focal lengths.

Still considering the optical path, downstream (Fig. 29, Fig. 33) of the collimator (48) there is a slit (64), which is used to reduce the "thickness" of the laser beam in the fast-axis direction without having to resort to, for example, bulkier inverted beam expanders or uncomfortably short focal collimator lenses. In the current implementation, with a 0.4- mm slit and a collimator (48) of relatively long focal length, by conscious choice, much of the overabundant light power emitted by the laser is "wasted" in favor of a simpler but equally effective optical configuration for the purposes of the intended measure.

Still considering the optical path, downstream (Fig. 29, Fig. 33) of the slit (64) there is the line-generation optics (52), similar to what was previously described in relation to the embodiment related to single line projection. The line-generating optics (52) can be a cylindrical lens that allows the laser beam to diverge in the slow-axis direction in such a way that a line is projected onto the surface of the powder level (1 ) where the measure takes place instead of a single dot. The beam divergence angle included by the line generation optics (52) is small, for example, between 0.5° and 2°, such as 1°. An inexpensive plane-concave lens with a focal length of -50 mm is used here, but we could also employ a Powell lens that would project a line with more uniform luminous intensity but is more expensive, a cylindrical plane-convex, biconcave or biconvex lens, or a diffractive line-generating optics.

Still considering the optical path, downstream (Fig. 29, Fig. 33) of the linegenerating optics (52) there is the previously described beam splitter (65) which is apt to split the single beam generated by the light source (50) into two mutually parallel and practically equal beams, although, as can be seen (Fig. 33) the optical paths are somewhat different. The beam splitter (65) can be made in the form of the component usually known as the lateral displacement beamsplitter. For example, in this embodiment, the distance between the two beams generated by separation from the beam splitter (65) is 24 mm. A greater distance improves the measure accuracy but increases the overall dimensions, and vice versa. In general, a useful range of the distance between the two generated beams between 10 mm and 40 mm is provided. It will be obvious to a person skilled in the field that other methods of splitting the beam could also be employed, but it is essential that the parity of the two beams be the same, otherwise, in the presence of fluctuations in the direction of the incoming beam in the plane parallel to the plane of plan representation (Fig. 33), the two output beams would fluctuate in opposite directions producing a measure error, thus being understandable the need to have a beam splitter (65) such that the condition of parity identity of the two beams generated by the splitter is maintained.

Still considering the optical path, downstream (Fig. 29, Fig. 33) of the beam splitter (65) there is the previously described perspective amplifier (66) which is apt to introduce the described convergence configuration between the two beams. In particular, the perspective amplifier (66) can be a Fresnel biprism whose function, precisely, is to amplify the perspective effect by converging the two originally parallel beams at one point, as previously described. The biprism employed has, for example, a vertex angle of 177° and is made of fused silica, including, in the orientation shown in the figure (Fig. 33), that is, with incidence on the flat base, a convergence angle (AC) of 1.41°, as calculable from Snell's law. The angle at the vertex can be made to include a convergence angle (AC) between 0.5° and 5°, preferably between 1° and 2°. It is convenient to define, for convenience of language in subsequent descriptions, as the "optical axis of the projector" the line bisecting the angle formed by the two converging laser beams as a result of passing through the perspective amplifier (66).

Still considering the optical path, downstream (Fig. 29, Fig. 33) of the perspective amplifier (66) there is an optical path splitter (67) which, as can be seen in the figures (Fig. 29, Fig. 33) is intended to overlap the optical axis of the camera (6) with the optical axis of the projection system (48, 50, 52, 64, 65, 66), so that the structured light illumination function of the light source (50) with projection of the two segments and the image acquisition function through the camera (6) share a same optical path within the light transmission system (12) for light guiding along the second axis (X) and, subsequently, for deviation along the first axis (W) of vision through the deviation system (11 ). In practice, the optical path splitter (67) is apt to overlap an optical axis of the camera (6) with an optical axis of the projection system (48, 50, 52, 64, 65, 66), such that the camera (6) and the projection system (48, 50, 52, 64, 65, 66) share a common optical path within the light transmission system (12) of the optical device (3). The optical path splitter (67) is essentially implemented by means of a second beam splitter different from the beam splitter (65) used to split the single beam generated by the light source (50) into two mutually parallel beams. The optical path splitter (67) is equipped with a semi-reflective surface that:

- in a first direction related to transmission, reflects the two laser beams from the light source (50) by transmitting them toward the second axis (X) within the light transmission system (12) such that they reach the deviation system (11) to be, then, deflected toward the mould (26) by projection of the two laser beams onto the powder level (L1 ) along the first axis (W) of vision; - in a second direction relative to reception, lets at least part of the light received from the vision field (22) pass without deviation within the mould (26), the light received from the vision field (22) being deflected by the deviation system (11 ) in such a way that it is deflected from the first axis (W) of vision toward the second axis (X) so that it reaches the camera (6) by passing through the optical path splitter (67) which lets the light pass along the second direction without deviation.

Thus, through the optical path splitter (67) (Fig. 29, Fig. 33) the optical path of the laser beams toward the vision field (22) and of the light coming from the vision field (22) which is in common within the light transmission system (12) is divided, the optical path being divided into a first optical sub-path related to the two laser beams coming from the light source (50) and oriented toward the vision field (22) and a second optical sub-path related to the light coming from the vision field (22) oriented toward the camera (6).

Thus, the camera (6) observes (Fig. 29, Fig. 33), through the semi-reflective surface of the optical path splitter (67), the image of the two light segments obtained by the projection of the laser beams deflected by the optical path splitter (67) itself. Since the optical path splitter (67), implemented by means of a second non-polarized beam splitter, transmits 50% of the incident light, only half of the scattered light from the surface of the powder layer (L1) is measured. Experimentally, it is observed that there is no need to exploit laser polarization to increase the efficiency of the system and, therefore, it is sufficient to use a non-polarized beam splitter for simplicity. Considering the light relative to the laser beams reflected by the optical path splitter (67), again there is a transmission of 50% of the light to the light-transmitting system (12), while the remaining 50% that is not reflected back to the light-transmitting system (12) must be absorbed, so as not to disturb the useful image that is acquired by the camera (6). For this reason, the optical path splitter (67) includes an appropriate blind surface blackening or anti-reflection treatment for subsequent absorption by a beam absorber or "beam-trap."

Between the camera (6) and the optical path splitter (67) there are (Fig. 29, Fig. 33) a filter (63) and a lens (7), as previously described.

The filter (63) is an interference filter, as previously explained in connection with the measures taken to avoid saturation of the camera (6) acquisition sensor.

The lens (7) is (Fig. 29, Fig. 33) preferably but not necessarily cautiously chosen with characteristics of low distortion and good resolution. The focal length is 30 mm and the resolution 160 Ip/mm, employed in conjunction with a camera (6) with 3.45 micrometer pixels. The aperture of the camera (6) results in an aperture ratio of F/16 with an acceptable depth of field. Alternatively, a 50-mm lens with a resolution of 100 Ip/mm could be used, for example, when coupled with a camera with 5.86-micrometer pixels. The lens (7) is focalized at about one-third of the measure range, i.e., the vision field (22), from its proximal end.

The camera (6) (Fig. 29, Fig. 33) may be, for example, of the type with a CMOS area sensor, preferably with a dynamic range of 72dB at 12bit, or, alternatively, with a dynamic range of 48dB at 8bit. The area of the camera sensor (6) needed for measure is a square or rectangle having a few hundred pixels per side, e.g., 200 pixels. Thus, a 1/3-inch sensor is more than sufficient. The CMOS sensor should preferably be the type equipped with global-shutter to avoid problems under dynamic conditions. To reduce the amount of heat-derived radiation collected and to prevent image blurring, the laser-type light source (50) is pulsed synchronously with the camera (6). In the simplest implementation, a camera (6) with a strobe output is used, with a remotely duration programmable pulse, which may range from 100 microseconds to 10 milliseconds, typically 1 millisecond. The pulse is emitted in synchrony with the integration period of the camera. The strobe signal drives a constant-current power pulsator (about 1 A) that powers the light source laser diode (50).

With reference to this further embodiment (Fig. 29, Fig. 33), the optic device (3) is advantageously realized differently from the other embodiments previously illustrated (Fig. 1 , Fig. 3, Fig. 4, Fig. 5, Fig. 6). In fact, while in the previous embodiments the optic device (3) comprises a case (9) containing a light transmission system (12) having lenses or optical fibres, in this further form of embodiment the optical device (3) comprises a case (9) containing a light transmission system (12) which consists of a tubular element devoid of optical components such as lenses or optical fibres in which the tubular element is internally blackened to avoid unwanted reflections on the inner walls.

Similar to the previously illustrated embodiments, the optical device (3) comprises (Fig. 29, Fig. 30, Fig. 31) a case (9) attached to the container (5) by means of a clamping system (8) of a known type, in which within the case (9) is housed a light deviation system (11 ) for light deviation with variation of the first axis (W) and transmission of light according to the second axis (X) of transmission. In this case, the light deviation system (11) can be realized in the form of a mirror, mounted at the first end (60) of the optic device (3). Preferably the light deviation system (11) is arranged so that the first axis (W) is inclined with respect to the second axis (X) by a second angle (AG2) between 70° and 140°, preferably between 80° and 95°, even more preferably 90°. In fact, in the case of moulds (26) that are too small, it may be difficult to be able to protrude far enough beyond (Fig. 18) the edge of the mould (26) with the optic device (3) of the vision device (1 ) without getting too close to the glowing snorkel (27). In such a situation it may be useful to orient the deviation system (11) so that the beam is deflected by, for example, 80°, instead of at a right angle. In this case it will obviously be necessary to correct the distance measure by the factor sin(80°)=0.985. The mirror is a mirror having temperature resistance features of at least 300°C without losing reflectance and without significant thermal deformation. Preferably it is envisaged, in contrast to the previously illustrated embodiments equipped with a pinhole lens, that the optic device (3) has a vision window equipped with a sapphire protection, resistant to scratches and molten steel splashes, mounted on the opening (10) with an inclination of 10° to avoid specular reflections towards the camera (6). The usable range in principle begins a few centimeters from the vision window and ends a few centimeters from the point of intersection of the projected beams, at the point where the algorithm can no longer distinguish between the images of the two nearly overlapping segments. Typically, the useful length of the vision field is about 250 mm.

In the further illustrated embodiment (Fig. 30, Fig. 32), a second heat sink (43") different from the first heat sink (43") previously described for cooling the camera (6) is also illustrated. The second heat sink (43') is disposed at least at one wall of the container (5), preferably at the bottom of the container (5), for cooling the entire assembly of components contained within the container (5) itself and is also to be considered applicable to the previously described embodiments (Fig. 17), as, to the present further embodiment (Fig. 30, Fig. 32) the first heat sink (43') is to be considered applicable. Similarly to what was explained with reference to the first heat sink (43'), the second heat sink (43') preferably includes internally a flow channel which is connected to the inlet (45) of the gaseous flow (44) and to the free space (13) for cooling the optic device (3). For example, the second heat sink (43") can be made in the form (Fig. 32) of a spirally wound pipe or having 180° bends to obtain mutually parallel sections defining said flow channel, the second heat sink (43") being able to be arranged on one or more walls of the container.

With reference to the beam splitter (65), it will be obvious to a person skilled in the field that multiple alternative embodiments to the one previously illustrated can be envisaged. There are, in fact, numerous variants that can perform the same functions performed by the prism assembly of the preferred embodiment, the choice being dictated only by the ease of construction and implementation or the ease of finding the components. For example, alternatives could be provided, such as:

- an embodiment (Fig. 35) in which, compared with the previously illustrated beam splitter (65), the semi-reflective and reflective surfaces of the beam splitter (65) are tilted in the opposite direction, achieving the same effect as a Fresnel biprism;

- an embodiment (Fig. 36) in which refraction is exploited on the exit faces instead of the reflective surfaces;

- an embodiment (Fig. 37) in which a first prismatic optical component (68) in the form of a beam splitter (65) of a simple type is used together with a second prismatic optical component (69) in the form of a right-angle prism, the first prismatic optical component (68) and the second prismatic optical component (69) being cemented stably on a common substrate (70) of the same material, with a low coefficient of thermal expansion, so that an almost monolithic assembly is obtained. In the figure (Fig. 37), the path of the first line or projection pattern (42') and the second line or projection pattern (42") is schematized without taking refractions into account, for illustrative simplicity;

- an embodiment (Fig. 38) in which a first prismatic optical component (68) is used in the form of a simple beam splitter along with a second prismatic optical component (69) in the form of a prism that are made as non-standard components, the first prismatic optical component (68) and the second prismatic optical component (69) being cemented stably onto a common substrate (70) of the same material, with a low coefficient of thermal expansion, so that an almost monolithic assembly is obtained. The use of nonstandard components is also possible due to the ease of design, all refractions being, by construction, at normal incidence.

Beam splitter of simple type is intended to denote the simplest embodiment among the many ones of beam splitters, in which the beam splitter consists of a cube made of two mutually bonded triangular prisms, devoid of birefringent materials by splitting light into rays by semi-reflective coating, regardless of their polarization, by splitting the received light in half between transmitted and reflected components.

The last two embodiments (Fig. 37, Fig. 38) illustrated allow (Fig. 39) advantageous operation of the optical path splitter (67), thus allowing the camera (6) to directly frame the vision field (22) as the light from the vision field is free to pass through the free space between the first prismatic optical component (68) and the second prismatic optical component (69), partly due to the fact that a small focal length and high diaphragm closure is used as explained above.

As previously explained, one can imagine other variations in relation to the shape of the first line or projection pattern (42') and the second line or projection pattern (42"), allowing for the projection of a more complex pattern but with a light beam converging at one point. For example, instead of a simple couple of segments, one could project a circumference from whose diameter one could deduce the projection distance. Such alternatives appear less advantageous because they would require the use of more complex and bulky optics, losing the essential advantage of being able to compact the measuring instrument.

With reference to the further embodiment comprising the projection of the first beaming line or pattern (42') and the second beaming line or pattern (42"), the camera (6) acquires a frame containing the image (Fig. 40) in black and white within which there are at least two acquired beaming lines or patterns in double projection (A', A") corresponding to the beaming lines or patterns (42', 42") generated and projected on the surface of the powder level (L1). In particular, in the case of the projection of two beaming lines (42', 42"), the camera (6) acquires, for each acquisition, a frame containing a first line acquired in double projection (A¹) corresponding to the acquisition of the first beaming line or pattern (42') and a second line acquired in double projection (A") corresponding to the acquisition of the second beaming line or pattern (42"). The acquired lines (A'. A") are essentially arranged along the columns of the acquired image and the objective is to identify for each line of the acquired image a couple of points identifying respectively the first acquired line in double projection (A') corresponding to the acquisition of the first line or projection pattern (42') and the second acquired line in double projection (A") corresponding to the acquisition of the second line or projection pattern (42"). If one observes the luminous intensity of a given line of the acquired image (Fig. 41) as a function of the columns, one can see that both the first line acquired in double projection (A') and the second line acquired in double projection (A") have a width of a few pixels. Various criteria can be used to identify the two characteristic points. First of all, one can identify the lines acquired in double projection (A', A") according to the methodology already explained above in relation to the identification of the lines acquired in single projection (A1 , A2, A3, An), with the only difference being that in this case on each acquired image two lines must be identified instead of a single line, being able to envisage, by analogy, filtering phases based on brightness, via application of a linear regression function.

The simplest method to identify the two characteristic points of the lines acquired in double projection (A¹, A") is that of the maximum, however a more effective method is based on the calculation of the barycentre relative to an appropriate interval containing the maximum of each intensity peak. For example, a procedure might be as follows: the maximum of the intensity peak is identified, then defined N points, such as three, four, five points, the barycentre is calculated according to the following formula: (11)

The value denoted by jmax represents the column index of the peak maximum (Fig. 41), aduij represents the intensity matrix, i.e. the frame or image acquired from the camera (6), and i represents the row index of the frame or image acquired.

In this way, the couple (i, xi) is identified, as exemplified in the figures (Fig. 41) where the result of applying formula (11) in relation to a given row of the acquired image is represented (Fig. 40). Repeating the procedure for each row of the frame or image acquired (Fig. 40) we obtain, therefore, a double series of points in which:

- the points of the first set are indicative of the positions of the barycentres of the luminous intensity values for each line of the frame or image acquired relative to the first line acquired in double projection (A¹) corresponding to the acquisition of the first beaming line or pattern (42')

- the points of the second set are indicative of the positions of the barycentres of the luminous intensity values for each line of the frame or image acquired relative to the second line acquired in double projection (A") corresponding to the acquisition of the second beaming line or pattern (42").

Subsequently, various strategies can be used to calculate (Fig. 42) the spacing measure (M) between the first line acquired in double projection (A^') and the second line acquired in double projection (A"). A consolidated method consists of performing a linear interpolation of the barycentres (Fig. 42), obtaining a first interpolated line representative of the first line acquired in double projection (A’) and a second interpolated line representative of the second line acquired in double projection (A"). The distance between the two lines representing the first line acquired in double projection (A^') and the second line acquired in double projection (A") is then calculated.

It should be noted that, due to mechanical construction, the first line or projection pattern (42') and the second line or projection pattern (42") are, in the case of a two-line projection, two parallel projected lines. However, the first interpolated line and the second interpolated line may not be parallel due to distortions introduced by irregularities in the surface of the powder layer (L1 ) such as localized accumulations or depressions. It is therefore necessary to define a procedure to calculate the spacing measure (M). Of the various methods, a particularly consolidated one is based on the following procedure:

- a first difference (Diffls) is calculated between the position of each barycentre of the first line acquired in double projection (A’), as calculated above, with respect to the second interpolated line;

- a second difference (Diff2B) is calculated between the position of each barycentre of the second line acquired in double projection (A"), as calculated previously, with respect to the first interpolated line;

- the mean value of the series of differences (Diffie, Diff2e) obtained is calculated and taken as the value of the distance between the first interpolated line and the second interpolated line;

- The measure of spacing (M) between the first line or projection pattern (42') and the second line or projection pattern (42") is calculated from the distance between the first interpolated line and the second interpolated line using the well-known formulae of Euclidean geometry, in a manner analogous to that set out with reference to the case of a single projected line, as will be evident to a person skilled in the field.

In the specific case of the frame (Fig. 40) under consideration, the spacing measure (M) between the first interpolated line and the second interpolated line calculated using the specified procedure is 112.2 pixels.

In order to obtain the position of the powder level (L1), i.e. the height (H1, H2, H3) of the vision device (1 ) relative to the position of the powder level (L1), it is sufficient to have a conversion map between the previously obtained measure of the distance between the first interpolated line and the second interpolated line expressed in pixels and the corresponding measure in millimetres. A conventional method consists of placing a reference surface at known distances from a representative sensor point, e.g. the nodal point of the lens, and noting down (Fig. 43) the relative distance in pixels between the first interpolated line and the second interpolated line. The series of positions of the reference surface must cover the entire measure range of the vision field (22) of the vision device (1 ).

As previously explained with reference to the single line projection embodiment, the vision device (1 ) is installed on the mould (26) of the casting machine (30), such that the vision device (1) is integral with the mould (26) and movable together with the mould (26) according to the direction of oscillation (53) of the mould (26). Consequently, since there is an oscillation of the vision device (1) with respect to the powder level (L1), usually with a known sinusoidal oscillation having a total amplitude generally of the order of a few millimetres, such as 7 mm, for example, the measure of the position of the powder surface consists of two elements, the first being the mean value of this oscillation and the second being the deviation due to the oscillatory motion. It will be evident that the methodology already explained with reference to the single line projection embodiment with calculation of sets of midpoints (P1, P2, P3, Pn) of the lines acquired in double projection for each of the lines or patterns of projection (42', 42") projected by the projection system (46, 50, 51) can be immediately extended to the present case.

The proposed method correctly identifies the position of the powder surface in a continuous casting mould. The measure accuracy is less than a millimetre and therefore adequate for the purpose.

Ultimately, the present invention relates to (Fig. 1, Fig. 2) a vision device (1 ) for (Fig. 18, Fig. 19) casting machine (30) having a mould (26) for casting a metal in a molten state, wherein the vision device (1) comprises (Fig. 2) an elaboration unit (40), an optic device (3) and an acquisition device (2) comprising a camera (6), wherein the optic device (3) and the acquisition device (2) are mutually arranged one after the other such that light from the optic device (3) is received by the camera (6). In other words, the optic device (3) is coupled to the acquisition device (2) in such a way as to create an optical path for the light that is directed from the optic device (3) towards the acquisition device (2) until it reaches the camera (6) for capturing images of the received light. The vision device (1) comprises (Fig. 18, Fig. 19) a fixing system (4) for installation on the casting machine (30), such that when the vision device (1) is in the installed condition an opening (10) of the optic device (3) is directed towards a surface of the metal layer in the molten state within the mould (26). In this manner such light from the mould (26) enters through the opening (10) according to a first axis (W) of vision. The optic device (3) has (Fig. 1 , Fig. 2) an oblong shape developing along a second axis (X) and the oblong shape is provided with a first end (60) and a second end (61) opposite the first end (60) with respect to said oblong shape. The light entry opening (10) is arranged at the first end (60) and the acquisition device (2) is arranged at the second end (61), such that the acquisition device (2) is positioned at a distance from the first end (60) for protection from heat and molten metal splashes. The optic device (3) comprises a light transmitting system (12) for guiding the light along the second axis (X) to the acquisition device (2) and a light deviation system (11) for deviation the light from the first axis (W) to the second axis (X) to the light transmitting system (12), such that when the vision device (1) is in an installed condition (Fig. 18, Fig. 19), the second axis (X) is arranged according to a first angle (AG1) between -20 and 20 degrees with respect to a horizontal plane (O), the first axis (W) being inclined with respect to the second axis (X) by a second angle (AG2) between 90 degrees and 150 degrees, preferably between 100 degrees and 140 degrees, the optimal value being approximately 120 degrees. In practice, the optic device (3) is configured for the deviation of light with variation of the first axis (W) of vision and transmission of light towards the acquisition device (2) according to a second axis (X) of transmission, so that the light hits the camera (6) of the acquisition device (2). Horizontal plane means a plane orthogonal to the direction of gravity.

The vision device (1) comprises a projection system (46, 48, 50, 51 , 52, 64, 65, 66, 67) of at least one beaming line or pattern (42, 42’, 42") of light for projection of the at least one beaming line or pattern (42, 42', 42") within the vision field (22) of the vision device (1) with obtainment of the at least one beaming line or pattern (42, 42', 42") positioned superficially on the metal casting powder cover in the molten state, the elaboration unit (40) being configured to acquire said series of images (11 , I2, I3) by means of the camera (6) wherein each image of the series of images (11 , I2, 13) comprises said at least one beaming line or pattern (42, 42', 42") disposed superficially on the metal casting powder cover in the molten state, the elaboration unit (40) being configured for processing said series of images (11, I2, I3) acquired by means of the camera (6) with obtainment of a measure of the position of said at least one line or projection pattern (42, 42', 42") disposed superficially on the cover of casting powder of the metal in the molten state, the elaboration unit (40) being configured for calculating a position of the powder level (L1) in the mould, the vision device (1) comprising means of communication for transmitting the measure of the powder level position (L1) in the mould. Preferably, the projection system (46, 48, 50, 51, 52, 64, 65, 66, 67) comprises at least one laser diode light source (50), a collimator (48) for focusing, a line generation optics (52) and an optical path variation system (11 , 47) for orienting the at least one beaming line or pattern (42, 42', 42") towards the molten metal casting powder cover.

The horizontal direction (O) is understood to be defined in relation to the vertical direction of gravity. The casting plane of the casting machine is understood to be arranged horizontally.

In the preferred embodiment of the present invention, the vision device (1 ) is installed on the mould (26) of the casting machine (30), such that the vision device (1) is integral with the mould (26) and movable along with the mould (26) according to the direction of oscillation (53) of the mould (26).

The present invention also relates to a method of detecting a level position measure (L1) of powder in a mould by means of a vision device (1) as described, wherein the detection method comprises the following steps, which are generally to be considered applicable either for the projection of a single beaming line (42) either for the projection of a projection pattern for simplicity as understood in the definition of a line, or for the projection of two beaming lines (42', 42") or two beaming patterns, or for the projection of more than two beaming lines or complex projection patterns such as grids or circles:

(a) projection of at least one beaming line or pattern (42, 42', 42") of light within the vision field (22) of the vision device (1 ) while obtaining at least one beaming line or pattern (42, 42', 42") positioned superficially on the cover of molten metal casting powder;

(b) acquiring a series of images (11, I2, I3, In) over time by means of the camera (6) of the vision device (1) wherein each image comprises the at least one beaming line or pattern (42, 42', 42") projected by the projection system (46, 48, 50, 51 , 52, 64, 65, 66, 67);

(c) analysis of each of the acquired images by application of a luminous intensity threshold with identification in the series of acquired images of a series of points of increased luminous intensity;

(d) for each of the images, application of a linear regression function with identification of acquired lines (Al, A2, A3, An, A', A"), wherein each acquired line corresponds to the acquisition of at least one line or pattern selected from one or more lines or patterns of said at least one line or pattern (42, 42', 42") in a different position of the vision device (1 ) with respect to the powder level (L1) wherein the different position is due to the oscillation along the oscillation direction (53) of the mould (26) and the vision device (1) fixed to the mould (26);

(e) for each of the acquired lines (A1 , A2, A3, An, A', A") calculation of a midpoint by obtaining a series of midpoints (P1, P2, P3, Pn) of the acquired lines (A1, A2, A3, An, A', A");

(f) calculation of the position of the powder level (L1 ) by applying trigonometric triangulation formulae or by means of a correspondence table to the series of midpoints (P1 , P2, P3, Pn).

In the specific case of projecting a first beaming line or pattern (42') and a second beaming line or pattern (42') said step (a) of projecting at least one beaming line (42, 42', 42') is a phase of projecting said first beaming line or pattern (42') and second beaming line or pattern (42'), said step (b) is a phase of acquiring a series of images (11 , I2, I3, In) in which each image comprises said first beaming line or pattern (42') and said second beaming line or pattern (42") said step (d) is a step of applying a linear regression function with identification of a first line acquired in double projection (A¹) corresponding to the acquisition of the first line or projection pattern (42') and a second line acquired in double projection (A") corresponding to the acquisition of the second line or projection pattern (42") for each of the series of images (11 , I2, I3, In) acquired, the step (f) of calculating the position of the powder level (L1) comprising a first sub-step of calculating a spacing measure (M) between the first line acquired in double projection (A¹) and the second line acquired in double projection (A") and a second sub-step of applying the trigonometric triangulation formulae or a correspondence table to obtain the measure of the position of the powder level (L1 ) in the mould.

The description of the present invention has been made with reference to the appended figures in a preferred embodiment thereof, but it is evident that many possible alterations, modifications and variations will be readily apparent to those skilled in the art in the light of the above description. Thus, it should be noted that the invention is not limited by the preceding description, but includes all such alterations, modifications and variations in accordance with the appended claims. USED NOMENCLATURE

With reference to the identification numbers in the attached figures, the following nomenclature was used:

1. Vision device

2. Acquisition device

3. Optic device

4. Fixing system

5. Contenitore - Container

6. Camera

7. Focus setting system

8. Clamping system

9. Case

10. Opening

11 . Deviation system

12. Transmission system

13. Free space

14. Pinhole optics

15. Mirror or prism

16. Lens

16’ First lens couple

16” Second lens couple

16”’ Third lens couple

16””. Fourth lens couple

17. Rod lens

18. Relay lens

19. Air gap

20. Optical fibre 21. Shell

22. Vision field

23. Light emitter

24. Light transmitter

25. Illumination field

26. Mould

27. Snorkel

28. Semi-finished product

29. Guiding system

30. Casting machine

31 . Meniscus

32. Powder supply system

33. Conduit

34. Spreader head

35. Tank

36. Pneumatic system

37. Control unit

38. First communication channel

39. Second communication channel

40. Elaboration unit

41 . Fixing system

42. Beaming line or pattern

42’. First beaming line or first beaming pattern

42”. Second beaming line or second beaming pattern

43’. First heat sink

43”. Second heat sink

44. Flow

45. Inlet 46. Projector

47. Reflective surface or prism

48. Collimator

49. Flow regulator

50. Light source

51. Light guide

52. Line generation optics

53. Oscillation direction

54. Chamber

55. O-Ring seat

56. Ring

57. Hole

58. Flange

59. Passage

60. First end

61. Second end

62. Level sensor

63. Filter

64. Slit

65. Beam-splitter

66. Perspective amplifier

67. Optical path splitter

68. First prismatic optical component

69. Second prismatic optical component

70. Substrate

A1. First acquired line in single projection

A2. Second acquired line in single projection

A3. Third acquired line in single projection A’. First acquired line in double projection

A”. Second acquired line in double projection

AC. Convergence angle

L1 . Powder level L2. Molten metal level

M. Spacing measure

ROI. Region of interest

X. Second axis

W. First axis

Claims

1. Vision device (1) for casting machine (30) of a melted-state metal provided with a mould (26), in which the vision device (1) includes a container (5), an elaboration unit (40), an optic device (3) and an acquisition device (2) provided with a camera (6), the optic device (3) and the acquisition device (2) being reciprocally placed one after the other in such a way that light from the optic device (3) is received by the camera (6) for acquisition of a series of images (11 , I2, I3), the vision device (1) comprising a fixing system (4) for installation on the casting machine (30) in such a way that, when the vision device (1) is in installed condition, an opening (10) of the optic device (3) is directed towards a surface of the melted-state metal level with casting powder cover inside the mould (26) in such a way that light coming from the mould (26) and placed inside a vision field (22) enters through the opening (10) according to a first axis (W), in which the optic device (3) has an oblong shape developing along a second axis (X) provided with a first end (60) and a second end (61 ) opposite to the first end (60), in which the opening (10) is placed at the first end (60) and in which the acquisition device (2) is placed at the second end (61) in such a way that the acquisition device (2) is placed in a spaced position with respect to the first end (60) for protection from heat and splashes of metal, the optic device (3) comprising a transmission system (12) of the light for the guide of the light along the second axis (X) to the acquisition device (2) and a deviation system (11) of the light for the deviation of the light from the first axis (W) to the second axis (X) in such a way that when the vision device (1 ) is in installed condition the second axis (X) is inclined by a first angle (AG1) between -20 and 20 degrees with respect to a horizontal-plane (O), the first axis (W) being inclined with respect to the second axis (X) by a second angle (AG2) between 90 degrees and 150 degrees, characterised in that the vision device (1 ) includes a projection system (46, 48, 50, 51 , 52, 64, 65, 66, 67) of at least one light beaming line or pattern (42, 42’, 42”) for projection of the at least one light beaming line or pattern (42, 42’, 42”) inside the vision field (22) of the vision device (1) with obtainment of at least one light beaming line or pattern (42, 42’, 42”) placed superficially on the casting powder cover of the melted-state metal, the elaboration unit (40) being configured for the acquisition of said series of images (11, I2, I3) by means of the camera (6) in which each image of the series of images (11 , I2, I3) includes said at least one light beaming line or pattern (42, 42’, 42”) superficially placed on the casting powder cover of the melted-state metal, the elaboration unit (40) being configured for the elaboration of said series of images (11 , I2, I3) acquired by the camera (6) with obtainment of a measure of the position of said beaming line (42) superficially placed on the casting powder cover of the melted-state metal, the elaboration unit (40) being configured for the calculation of a position of the level (L1) of the powder in the mould, the vision device (1) comprising communication means for transmission of the measure of the position of the level (L1 ) of the powder in the mould.

2. Vision device (1) according to the previous claim characterised in that the vision device (1) is installed on the mould (26) of the casting machine (30) in such a way that the vision device (1 ) is integral with the mould (26) and movable with the mould (26) according to the oscillation direction (53) of the mould (26).

3. Vision device (1) according to any of the previous claims characterised in that the container (5) contains at least:

- a focus setting system (7) for the focusing of images of the light received by the optic device (3) in such a way that the light is focalized at a focusing point placed along the second axis (X) at the opposite side with respect to the side on which the optical device (3) is present;

- said camera (6) placed in such a way that an acquisition sensor is placed in the focusing point, for acquisition of images of the light received by the optic device (3) and focalized by the focus setting system (7).

4. Vision device (1) according to any of the previous claims characterised in that the deviation system (11 ) of the light and the transmission system (12) of the light are housed inside a shell (21) which is placed inside a case (9), between the shell (21 ) and the case (9) being present a free space (13) for introduction of a gaseous flow (44) for cooling and cleaning, the free space (13) constituting a path for the gaseous flow (44) towards an exit of the gaseous flow (44) consisting of the opening (10).

5. Vision device (1) according to the previous claim characterised in that the vision device (1) includes a distribution system (54, 55, 58) placed in correspondence of a coupling interface between the acquisition device (2) and the optic device (3), the distribution system (54, 55, 58) comprising an insulation flange (58) between the acquisition device (2) and the optic device (3) and a collar (56) spaced with respect to the flange (58) with formation of a chamber (54) of distribution of the gaseous flow (44), the chamber (54) being provided with passages (59) for introduction of the gaseous flow (44) from the acquisition device (2), the collar (56) being provided with radial holes (57) of distribution of the gaseous flow (44) for delivery of the gaseous flow (44) from the chamber (54) inside the free space (13) of the optic device (3).

6. Vision device (1) according to any of the previous claims 4 to 5 characterised in that

It includes at least one heat sink (43’, 43”) selected between:

- a first heat sink (43’) which is applied on the camera (6) in such a way that heat coming from the camera (6) is transferred to the first heat sink (43’), the first heat sink (43’) internally comprising a respective flow channel of the first heat sink (43’) for the passage of the gaseous flow (44);

- a second heat sink (43”) which is applied in correspondence of at least one wall of the container (5), for the cooling of a series of components contained into the container (5), the second heat sink (43”) internally comprising a respective flow channel of the second heat sink (43”) for the passage of the gaseous flow (44).

7. Vision device (1) according to any of the previous claims characterised in that the at least one light beaming line or pattern (42, 42’, 42”) is selected between a projection pattern in the form of a segment, a projection pattern in the form of a grid, a projection pattern in the form of a circumference.

8. Vision device (1) according to any of the previous claims characterised in that the at least one light beaming line or pattern (42, 42’, 42”) consists of light having wavelength between 400 nm and 480 nm.

9. Vision device (1) according to any of the previous claims characterised in that the projection system (46, 48, 50, 51 , 52, 64, 65, 66, 67) includes at least one light source (50) with a laser diode, a collimator (48) for focusing, a line generation optics (52) and an optical path variation system (11 , 47) for orientation of the at least one light beaming line or pattern (42, 42’, 42”) towards the casting powder cover of the melted-state metal.

10. Vision device (1) according to any of the previous claims characterised in that the deviation system (11) of the light includes a pinhole optics (14) combined to a mirror or a prism (15) for deviation of the light, the pinhole optics (14) being made in form of optics lenses in which a diaphragm for setting the opening of the optics lenses has a hole diameter of the diaphragm between 0.5 mm and 8 mm.

11 . Vision device (1) according to the previous claim characterised in that the pinhole optics (14) implements said vision field (22) which is expanded in a direction oriented according to the first axis (W) in which the vision field (22) corresponds to a vision cone with aperture angle between 10 and 150 degrees.

12. Vision device (1) according to any of the previous claims characterised in that the transmission system (12) of the light includes one or more optical elements selected from:

- group of achromatic spherical lenses (16) placed in pairs with obtainment of achromatic doublets for the guide of the light along the transmission system (12).;

- GRIN type lenses (18) which are placed along the second axis (X) and spaced by airgaps (19).

- rod lenses (17) or Hopkins lenses;

- a coherent glass optical fibres bundle in such a way that the transmission system (12) is flexible.

13. Vision device (1) according to any of the previous claims characterised in that the vision device (1) includes a lighting system (23, 24) comprising a light emitter (23) with obtainment of an illuminations field (25) at least partially superimposed to the vision field (22) of the vision device (1).

14. Vision device (1) according to any of the previous claims characterised in that the projection system (46, 50, 51 ) is placed externally with respect to the vision device (1 ).

15. Vision device (1) according to any of the previous claims 1 to 13 characterised in that the projection system (46, 48, 50, 51, 52, 64, 65, 66, 67) is placed inside the container (5).

16. Vision device (1) according to the previous claim characterised in that it includes an optical path splitter (67) for overlapping of an optical axis of the camera (6) with an optical axis of the projection system (46, 48, 50, 51 , 52, 64, 65, 66, 67), in such a way that the camera (6) and the projection system (46, 48, 50, 51, 52, 64, 65, 66, 67) share a common optical path inside the transmission system (12) of the light of the optic device (3).

17. Vision device (1) according to any of the previous claims 15 to 16 characterised in that the transmission system (12) of the light comprises a tubular element without optical components like lens or optical fibres, in which the tubular element is internally blackened to avoid light reflections on the internal walls.

18. Vision device (1) according to any of the previous claims characterised in that the projection system (46, 48, 50, 51, 52, 64, 65, 66, 67) is configured for the projection of a single light beaming line or pattern (42, 42’, 42”) inside the vision field (22) of the vision device (1).

19. Vision device (1) according to any of the previous claims 1 to 17 characterised in that the projection system (46, 48, 50, 51, 52, 64, 65, 66, 67) is configured for the projection of two of said at least one light beaming line or pattern (42, 42’, 42”) for the projection of a first light beaming line or pattern (42’) and of a second light beaming line or pattern (42”) inside the vision field (22) of the vision device (1) with obtainment of two light beaming lines or patterns (42’, 42”) placed superficially on the casting powder cover of the melted-state metal.

20. Vision device (1) according to the previous claim characterised in that the projection system (46, 48, 50, 51 , 52, 64, 65, 66, 67) includes a beam-splitter (65) for the split of a single beam generated by the light source (50) into two reciprocally parallel beams.

21 . Vision device (1) according to the previous claim characterised in that the beam-splitter (65) is selected between a lateral displacement beam-splitter, a beam-splitter consisting of a first prismatic optical component (68) in the form of a beam-splitter together with a second prismatic optical component (69) in the form of a right-angle prism in which the first prismatic optical component (68) and the second prismatic optical component (69) are cemented on a common substrate (70) made of the same material of the two prismatic optical components (68, 69) with obtainment of a monolithic assembly.

22. Vision device (1) according to any of the previous claims 19 to 21 characterised in that the projection system (46, 48, 50, 51 , 52, 64, 65, 66, 67) includes a perspective amplifier (66) for obtainment of a reciprocal convergence condition according to a convergence angle (AG) between the two reciprocally parallel beams obtained by means of said beam-splitter (65).

23. Vision device (1) according to the previous claim characterised in that the convergence angle (AG) is between 0.5° and 5°.

24. Vision device (1 ) according to any of the previous claims 22 to 23 characterised in that the perspective amplifier (66) is a Fresnel biprism.

25. Powder supply system (32) for mould (26) for distribution of casting powder inside the mould (26), in which the powder supply system (32) includes a control unit (37), a container (35) and a conduit (33) for the feed of a mixture of transport air and casting powder to one or more spreader heads (34) of the casting powder towards the mould (26) characterised in that the powder supply system (32) includes a vision device (1) according to the previous claim, the elaboration unit (40) being configured for the transmission of the measure of the position of the level (L1) of the powder in the mould to the control unit (37) by means of a first communication channel (38) of the communication means, the control unit (37) being configured for the reception of the measure of the position of the level (L1 ) of the powder in the mould and being configured for the reception of a measure of the position of the steel level in the mould from a level sensor (62), the control unit (37) being further configured for the calculation of a measure of the thickness of the mould casting powder in the mould on the basis of measure of the position of the level (L1 ) of the powder in the mould and measure of the position of the steel level, the control unit (37) being configured for the calculation of an adjustment signal for generation of a powder feeding control from the container (35) to the mould.

26. Detection method of a position measure of level (L1 ) of powder in mould by means of a vision device (1 ) according to any of the previous claims 1 to 24 in which the detection method includes the following steps: (a) projection of at least one light beaming line or pattern (42, 42’, 42”) inside the vision filed (22) of the vision device (1) with obtainment of at least one light beaming line or pattern (42, 42’, 42”) placed superficially on the casting powder cover of the melted-state metal;

(b) acquisition of a series of images (11 , I2, I3, In) in the course of time by the camera

(6) of the vision device (1) in which each image includes the at least one light beaming line or pattern (42, 42', 42”) projected by the projection system (46, 48, 50, 51, 52, 64, 65, 66, 67);

(c) analysis of each of the acquired images by means of application of a threshold of luminous intensity with identification in the series of acquired images of a series of points with greater brightness intensity;

(d) for each of the images application of a linear regression function with identification of acquired lines (A1 , A2, A3, An, A’, A”), in which each acquired line corresponds to the acquisition of at least one light beaming line or pattern selected between one or more lines or patterns of said at least one light beaming line or pattern (42, 42’, 42”) in a different position of the vision device (1) with respect to the level of the powder (L1) in which the different position is due to the oscillation along the oscillation direction (53) of the mould (26) and of the vision device (1) fixed to the mould (26);

(e) for each of the acquired lines (A1 , A2, A3, An, A’, A”) calculation of a midpoint with obtainment of a series of average points (P1 , P2, P3, Pn) of the acquired lines (A1 , A2, A3, An, A’, A”);

(f) calculation of the position of the level (L1 ) of the powder by means of application to the series of average points (P1 , P2, P3, Pn) of trigonometric formulas of triangulation or by means of a correspondence table.

27. Detection method of a position measure of level (L1) of powder in mould according to the previous claim, in which the vision device (1) is made according to claim 19 characterized in that: said phase (a) of projection of at least one light beaming line or pattern (42, 42’, 42”) is a projection phase of said first light beaming line or pattern (42’) and second light beaming line or pattern (42”), said phase (b) is an acquisition phase of a series of images (11 , I2, I3, In) in which each image includes the first light beaming line or pattern (42’) and the second light beaming line or pattern (42”), said phase (d) is a phase of application of a linear regression function with identification of a first acquired line in double projection (A’) corresponding to the first light beaming line or pattern (42’) and a second acquired line in double projection (A”) corresponding to the second light beaming line or pattern (42”) for each of the series of acquired images (11, I2, I3, In), the phase (f) of calculation of the position of the level (L1 ) of the powder comprising a first sub-phase of calculation of a spacing measure (M) between the first acquired line in double projection (A’) and the second acquired line in double projection (A”) and a second sub-phase of application of trigonometric formulas of triangulation or of a correspondence table with obtainment of the measure of the position of the level (L1 ) of the powder in the mould.