WO2025081283A1

WO2025081283A1 - Optical monitoring of the viscoelasticity of a milk curd in a milk product manufacturing process

Info

Publication number: WO2025081283A1
Application number: PCT/CA2024/051382
Authority: WO
Inventors: Daniel Cote
Original assignee: Bliqc Inc
Current assignee: Bliqc Inc
Priority date: 2023-10-19
Filing date: 2024-10-18
Publication date: 2025-04-24
Anticipated expiration: 2026-04-19

Abstract

Methods for determining a set-time and/or a suggested cut-time for a milk curd in a milk product manufacturing process are provided. A coherent light beam is projected on the milk curd, and a viscoelasticity of the milk curd is monitored. The monitoring includes obtaining, over time, a plurality of light scattering data sets each representative of a scattering of the coherent light beam in the milk curd over an exposure period; in real time, calculating a speckle contrast parameter from the light scattering data set; and; monitoring the evolution of the speckle contrast parameter over time. The set-time and/or suggested cut-time for the milk curd is determined based on the evolution of the speckle contrast parameter. An optical system for monitoring a viscoelasticity of a milk curd in a milk product manufacturing process is also provided.

Description

OPTICAL MONITORING OF THE VISCOELASTICITY OF A MILK CURD IN A MILK PRODUCT MANUFACTURING PROCESS

TECHNICAL FIELD

The technical field generally relates to methods and systems monitoring the viscoelasticity of a milk curd to assist in the manufacturing process of a cheese product or other milk product, and in particular concerns using such monitoring to determine a set-time and/or cut-time of the milk curd.

BACKGROUND

Typical cheese manufacturing processes involve a well-known series of steps to transform a starting milk substance into a milk curd, which then stands still for the curdling process to take place, a phase that can take a variable amount of time. At one moment the milk suddenly starts to coagulates or curdles, and the firmness of the milk curd changes rapidly, marking the onset of a coagulation phase. This transition moment is designated as the set-time of the milk curd. After an additional period of time, the milk curd is then cut in order to drain whey out of the curd. Typically, cheesemakers will choose a cut-time based on individual experience and/or predetermined recipes. Some cheesemakers choose to cut the curd early rather than late as a rule, as knives or curd breakers are known to crush the curd rather than cutting it cleanly if the coagulum (i.e. the coagulated milk curd) becomes too firm. Cutting when the coagulum is too firm also retards syneresis and results in high-moisture cheese. However, cutting when the coagulum is too soft also has its drawbacks, as it may decrease cheese yield due to increased loss of fat and curd fines. Thus, a correct selection of the cut-time helps maximize yield, increase the quality and improve the homogeneity of cheese.

In view of these issues, it would be helpful to cheese manufacturing processes to benefit from reliable approaches to determine the set-time of a milk curd, its cuttime, or both. SUMMARY

In accordance with one aspect, there is provided a method for determining a suggested cut-time for a milk curd in a milk product manufacturing process. The method comprises: a. projecting a coherent light beam on the milk curd; b. monitoring a viscoelasticity of the milk curd, said monitoring comprising: i. obtaining, over time, a plurality of light scattering data sets each representative of a scattering of the coherent light beam in the milk curd over an exposure period; ii. in real time, calculating a speckle contrast parameter from the light scattering data set; and iii. monitoring the evolution of the speckle contrast parameter over time; and c. determining the suggested cut-time for the milk curd based on the evolution of the speckle contrast parameter.

In some implementations, the milk product is a cheese product.

In some implementations, the plurality of light scattering data sets comprises a plurality of 2D images, the exposure period for each 2D image having a span of the order of a millisecond or shorter, each 2D image comprising a speckle pattern. Calculating the speckle contrast parameter for each 2D image may include: a. dividing the 2D image into a plurality of tiles; b. for each tile, calculating a speckle contrast value; and c. calculating an image average of the speckle contrast values for the plurality of tiles of said 2D image, and outputting said image average as the speckle contrast parameter for said image.

In some implementations, obtaining a plurality of light scattering data sets comprises measuring, at different time intervals, a plurality of light intensity signals, the exposure period for each light intensity signal having a length sufficient to contain a variation in a speckle contribution. Calculating the speckle contrast parameter for each light intensity signal may comprise calculating a temporal contrast value over the corresponding exposure period, and outputting said temporal contrast value as the speckle contrast parameter for said light intensity signal.

In some implementations, determining the suggested cut-time for the milk curd comprises: a. identifying a set-time corresponding to an onset of a coagulation phase in said milk curd based on a change in the evolution of the speckle contrast parameter; and b. determining the suggested cut-time for the milk curd based on the settime and on at least one pre-selected condition.

The at least one pre-selected condition on which the determining of the suggested cut-time may be based comprises a predetermined delay between the set-time and the suggested cut-time. The method may further comprise continuing the monitoring of step b. past the set-time, and the at least one pre-selected condition on which the determining of the suggested cut-time is based may comprise an observation of a further change in the evolution of the speckle contrast parameter after set set-time.

In some implementations, the method further comprises an additional step of signaling an inception of the suggested set-time to an operator of the milk product manufacturing process.

In accordance with another aspect, there is provided a method for determining the onset of a coagulation phase in a milk curd in a milk product manufacturing process, comprising: a. projecting a coherent light beam on the milk curd; and b. during a pre-coagulation phase of said milk product manufacturing process, monitoring a viscoelasticity of the milk curd, said monitoring comprising: i. obtaining, over time, a plurality of light scattering data sets, each representative of a scattering of the coherent light beam in the milk curd over an exposure period; ii. in real time, calculating a speckle contrast parameter from each light scattering data set; and iii. monitoring the evolution of the speckle contrast parameter over time; whereby a moment at which a change in the evolution of the speckle contrast parameter is observed corresponds to the onset of the coagulation phase.

In some implementations, the plurality of light scattering data sets comprises a plurality of 2D images, the exposure period for each 2D image having a span of the order of a millisecond or shorter, each 2D image comprising a speckle pattern. Calculating the speckle contrast parameter for each 2D image may comprise: a. dividing the 2D image into a plurality of tiles; b. for each tile, calculating a speckle contrast value; and c. calculating an image average of the speckle contrast values for the plurality of tiles of said 2D image, and outputting said image average as the speckle contrast parameter for said image.

In some implementations, milk product is a cheese product.

In accordance with another aspect, there is provided an optical system for monitoring a viscoelasticity of a milk curd in a milk product manufacturing process, comprising:

- a coherent light source generating a coherent light beam and configured to project said coherent light beam on the milk curd, a scattering of the coherent light beam in the milk curd resulting in the emission of a scattered light beam from the milk curd;

- a light sensing device configured to detect the scattered light beam;

- a controller operatively connected to the light detecting device and configured to obtain therefrom, over time, a plurality of light scattering data sets each representative of the scattering of the coherent light beam in the milk curd over an exposure period;

- a processing unit configured to: i. in real time, calculate a speckle contrast parameter from each light scattering data set; ii. monitor the evolution of the speckle contrast parameter over time; and iii. determine a suggested cut-time for the milk curd based on the evolution of the speckle contrast parameter.

In some implementations, the coherent light source comprises a laser diode.

In some implementations, the light sensing device comprises a 2D matrix of pixels.

In some implementations, the light sensing device comprises a photodiode. Other features and advantages will be better understood upon of reading of detailed embodiments with reference to the appended drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a graph showing the evolution of the speckle contrast parameter of a milk curd over time.

FIG. 2 is a flow chart of a method for determining a cut-time for a milk curd in a milk product manufacturing process according to one embodiment.

FIG. 3 is a schematic representation of an optical system for monitoring a viscoelasticity of a milk curd in a milk product manufacturing process according to one embodiment.

FIG. 4 is a flow chart of a spatial analysis for calculating the speckle contrast parameter according to one variant.

FIGs. 5A, 5B and 5C are three examples of 2D images acquired towards the beginning, the middle and the end of the pre-coagulation phase, respectively.

FIG. 6 shows the zone of an image selected for the purpose of calculating a speckle contrast value; FIG. 6A shows the division of the selected zone into tiles.

FIG. 7 is a flow chart of a temporal analysis for calculating the speckle contrast parameter according to one variant.

DETAILED DESCRIPTION

In the following description, similar features in the drawings have been given similar reference numerals. In order not to unduly encumber the figures, some elements may not be indicated on some figures if they were already mentioned in preceding figures. It should also be understood herein that the elements of the drawings are not necessarily drawn to scale and that the emphasis is instead being placed upon clearly illustrating the elements and structures of the present embodiments.

The terms “a”, “an” and “one” are defined herein to mean “at least one”, that is, these terms do not exclude a plural number of items, unless stated otherwise. Terms such as “substantially”, “generally” and “about”, that modify a value, condition or characteristic of a feature of an exemplary embodiment, should be understood to mean that the value, condition or characteristic is defined within tolerances that are acceptable for the proper operation of this exemplary embodiment for its intended application.

Unless stated otherwise, the terms “connected” and “coupled”, and derivatives and variants thereof, refer herein to any structural or functional connection or coupling, either direct or indirect, between two or more elements. For example, the connection or coupling between the elements may be mechanical, optical, electrical, logical, or any combination thereof.

In the present description, the terms “light” and “optical”, and variants and derivatives thereof, are used to refer to radiation in any appropriate region of the electromagnetic spectrum. The terms “light” and “optical” are therefore not limited to visible light, but can also include, without being limited to, the infrared or ultraviolet regions of the electromagnetic spectrum. Also, the skilled person will appreciate that the definition of the ultraviolet, visible and infrared ranges in terms of spectral ranges, as well as the dividing lines between them, may vary depending on the technical field or the definitions under consideration, and are not meant to limit the scope of applications of the present techniques.

To provide a more concise description, some of the quantitative expressions given herein may be qualified with the term "about". It is understood that whether the term "about" is used explicitly or not, every quantity given herein is meant to refer to an actual given value, and it is also meant to refer to the approximation to such given value that would reasonably be inferred based on the ordinary skill in the art, including approximations due to the experimental and/or measurement conditions for such given value.

In the present description, the term “about” means within an acceptable error range for the particular value as determined by one of ordinary skill in the art, which will depend in part on how the value is measured or determined, i.e. the limitations of the measurement system. It is commonly accepted that a 10% precision measure is acceptable and encompasses the term “about”.

In the present description, when a broad range of numerical values is provided, any possible narrower range within the boundaries of the broader range is also contemplated. For example, if a broad range value of from 0 to 1000 is provided, any narrower range between 0 and 1000 is also contemplated. If a broad range value of from 0 to 1 is mentioned, any narrower range between 0 and 1 , i.e. with decimal value, is also contemplated.

In accordance with various aspects, there are provided methods and an optical systems monitoring the viscoelasticity of a milk curd to assist in a cheese manufacturing process. Advantageously, in some implementations, the provided methods are carried out in real-time and directly in the milk curd, without the need for sampling.

Viscoelasticity can be understood as the relationship between the viscous, or liquid-like, and elastic, or solid-like, behavior of a substance. In some implementations, the monitoring of the viscoelasticity may entail a determination of the firmness of the substance, as for example the case in a milk curd of a cheese manufacturing process. As will be explained below, viscoelasticity can be related to a speckle contrast parameter. In some implementations, the optical methods and systems described herein may be applied in the context of a milk-product manufacturing process, for example a cheese manufacturing process. Although the description is framed in the context of cheese manufacturing, it will be readily understood by one skilled in the art that the described method and system may alternatively be applied to other milk products.

Cheese manufacturing process

Typical cheese manufacturing processes first involve heating a starting milk substance and reducing its pH. A culture of bacteria, referred to in the art as the rennet, is then added, and the milk is stirred for a very short time to ensure that the culture is evenly distributed in the milk. The resulting substance stands still for the curdling process to take place. As part of this process, the enzymatic culture firstly breaks down the envelope of the milk casein which can take a variable amount of time, for example around 5 to 6.5 minutes. This period of time may be referred to in the art as the enzymatic phase, pre-coagulation phase or as the setting period. During the enzymatic phase, the texture of the milk curd undergoes some changes, but these changes are typically minor and do not substantially affect the overall properties of the milk curd. However, at one moment the milk suddenly starts to coagulate, and the viscoelasticity and firmness of the milk curd changes rapidly, marking the onset of a coagulation phase. This transition moment between the end of the enzymatic phase and the onset of the coagulation phase is designated as the set-time of the milk curd. The set-time of the milk curd is often considered an important information for cheesemakers as it is traditionally used as a reference to determine the cut-time, that is, the moment at which the milk curd is cut in order to drain the whey. Typically, cheesemakers will choose a cut-time based on individual experience and/or predetermined recipes, such as a multiplicative factor of the settime. Some cheesemakers choose to cut the curd early rather than late as a rule, as knives or curd breakers are known to crush the curd rather than cutting it cleanly if the coagulum (i.e. the coagulated milk curd) becomes too firm. Cutting when the coagulum is too firm also retards syneresis and results in high-moisture cheese. However, cutting when the coagulum is too soft also has its drawbacks, as it may decreases cheese yield due to increased loss of fat and curd fines. Thus, a correct determination of the set-time is key to determining and suggesting an optimal cuttime, which in turn helps maximize yield, increase the quality and improve the homogeneity of cheese.

FIG. 1 is an example graph of the evolution of the viscoelasticity of a milk curd as a function of time, obtained by the inventors in a test setting using the method described herein, based on a speckle contrast parameter as explained below. A traditional set-time choice could for example be around 25 minutes from the beginning of the manufacturing process (time 0 on the graph, corresponding to the introduction of the rennet). Too short or too long of a delay between the set-time and the cut-time changes the texture of the recipe, as well as the amount of water and fat retained by the curds. Preselected cut-time values calculated from the beginning of the manufacturing process disregard the effects of variations in the speed of the enzymatic mechanisms at play and impacts the quality of the results. Firstly, the set-time and the length of the coagulation phase are never precisely known, and depend on conditions, seasons, etc. Secondly, the cheesemaker should avoid touching the milk curd as much as possible, as even minor disturbances can dramatically affect the curdling process. However, cheesemakers still typically put a hand into the milk curd to check that the curd has a "clean cut”. Using the method or system of the present invention, the cheese maker can me notified earlier in the process that a suggested cut-time has been reach, enabling the confident cutting of the milk curd at a moment closer to an optimal moment, leading to better quality results.

In some embodiments, there is provided a method for determining a suggested cut-time in a milk product manufacturing process, such as a cheese product manufacturing process. Referring to FIGs. 2 and 3, embodiments of the method 100 described herein and of a corresponding monitoring system 20 are schematically illustrated.

The method 100 first includes projecting 102 a coherent light beam 22 on the milk curd 24. The coherent light beam 22 may be embodied by any beam of light exhibiting spatial and/or temporal coherence. As understood by one skilled in the art, a coherent light beam may have a spatial or temporal coherence; spatial coherence refers to a fixed phase relationship between respective electric field values at a same moment in different locations, and temporal coherence a fixed phase relationship between respective electric field values at a same location but at different times. In some implementations, the coherent light beam is a laser beam. The coherent light beam may be generated by a coherent light source 21. The coherent light source 21 may for example be embodied by a laser source based on any type of gain medium such as for example a gas laser, a solid-state laser, a fiber laser, a dye laser or a semiconductor laser or laser diode. In some implementations, the coherent light source is embodied by a laser diode, such as for example a Fabry-Perot (FP) laser diode, a Distributed Feedback (DFB laser diode), a Distributed Bragg Reflector (DBR) laser diode, a Quantum Cascade (QCL) laser diode, Vertical Cavity Surface Emitting Lasers (VCSELs) laser diode, or the like. In some variants, the laser diode may be a spatial multimode laser diode. In other variants, the laser diode may be a singlemode laser diode.

The method 100 next includes monitoring 104 a viscoelasticity of the milk curd 24 during a pre-coagulation phase of the milk product manufacturing process, that is, during the phase between the addition of the rennet and the set-time, corresponding to the period between 0 and 6 minutes in the example of FIG. 1. This monitoring involves obtaining 106, over time, a plurality of light scattering data sets each representative a scattering of the coherent light beam 22 in the milk curd 24 over an exposure period. In the configuration shown in FIG. 3, the coherent light beam 22 is projected on the milk curd 24 at a non-zero incidence angle 0. A fraction of the light from the coherent light beam 22 undergoes specular reflection at the surface 25 of the milk curd 24 and leads to a reflected light beam 22’ at a reflection angle 0’ equal and opposite to the incidence angle 0. Another fraction of the coherent light beam 22 is transmitted through the surface 25 of the milk curd 24 and is elastically scattered by the particles inside the milk curd 24. Scattered light then escapes the milk curd 24 through the milk curd surface 25 in random directions. The optical system 20 includes a light sensing device 26 configured to detect a scattering light beam 28. In different implementations, the light sensing device 26 may be embodied by a camera, a photodiode, or the like. The light sensing device 26 is positioned over the milk curd 24 to detect the scattered light in its field of view, which defines the scattered light beam 28, thereby obtained the light scattering data sets.

The monitoring 104 of a viscoelasticity of the milk curd further includes calculating 108 a speckle contrast parameter from the light scattering data sets.

The term “speckle” is used in the art to designate random granular patterns which can be observed upon imaging the projection of a coherent light beam on a rough surface. The appearance of the light reflected by such a surface typically includes variations in intensity, or bright and dark “specks” in the reflected pattern. The speckle effect is a result of the constructive and destructive interference of light waves that take place when coherent light, such as a laser beam, is used as the light source. Speckles are therefore known to arise from the interference of a coherent light beam with itself after it has interacted with an irregular surface or a scattering volume. Theoretically, a perfectly monochromatic, perfectly coherent laser beam forms a wave having a planar wavefront, which would lead to a perfectly uniform image on a light sensor. However, when such a wave hits a scattering medium, either in reflection or transmission, the irregularities of the scattering change the wavefront, resulting in an irregular intensity pattern in the reflected or transmitted light. The random granular patterns in the intensity pattern are known in the art as speckles. A perfectly reflective or perfectly flat surface would not generate speckles. It is known in the art to polished surfaces of mirrors, lenses or microscope slides with variabilities lower than 1/1 Oth of the wavelength to avoid speckle generation. A speckle pattern may be observed even from a flat surface if light is allowed to transmit through the surface is then backscattered and transmitted back though the surface (a diffuse reflection).

Different parameters of patterns of speckles such as the average intensity, the difference between the most intense and least intense speckles, changes over time, etc. are all indicators of surface roughness. This technique is for example used to measure metal surfaces in the industrial field.

In the context of the method described herein, the scattered light detected by the light sensing device 26 is not a result of specular reflection of the coherent light beam on an irregular surface, but results from scattering of the coherent light transmitted through the surface of the milk curd into the milk immediately below the surface 25. The particles inside the milk curd, irrespectively of their nature, disturb the wavefront of the coherent light beam and scatter light waves randomly towards the surface 25 of the milk curd 24. The disturbances induced in the wavefront lead to interference patterns on the light sensing device 26, producing speckles. It will be noted that in this context the speckles are not caused by roughness on the surface 25 of the milk curd: as a matter of fact, the liquid surface 25 is usually perfectly smooth. The speckles appear because light has penetrated a few millimetres into the liquid of the milk curd 24, and the return of this disturbed light to the light, coming from different places in the milk curd 24, that creates the speckles. But just as the roughness of the surface gives the appearance of speckles, here the random volumetric perturbations will also give the appearance of speckles, and the evolution of the scattering during the measurement will ultimately affect the contrast. The Brownian motion of the scatterers in the medium are responsible for affecting the contrast by dynamically changing the speckle pattern. The change of speckle pattern is an indicator of this Brownian motion, which is related to viscoelastic properties of the medium.

The speckle contrast parameter may be embodied by any measurable or calculated variable having a value scaling with the level of contrast in the speckles observed from the scattering of the coherent light beam. Non-limitative examples of speckle contrast parameters are provided below.

As will be readily understood by one skilled in the art, the exposure period, also referred to as the exposure time, corresponds to the length of time over which photons scattered by the milk curd are collected by the light sensing device for a given light scattering data set. In some implementations, the duration of the exposure period is selected so as to enable the detection of variations in the observed speckles. Speckle patterns evolve in both time and space at an evolution speed dependent on the viscoelasticity of the observed medium. For a given data point, the length of the exposure period is preferably commensurate with a small number of appearance and disappearance cycles of speckles in the observed area on the medium. Selecting an exposure time which is either too short or too long may lead to a randomizing or an averaging effect masking the time evolution of the speckle contrast. In some embodiments, it has been found by the inventor that an exposure period of the order of a millisecond may be suitable in the context of a cheese manufacturing process.

The monitoring 104 a viscoelasticity of the milk curd further includes monitoring 110 the evolution of the speckle contrast parameter over time. In some embodiments, the method may include a step of identifying 140 the set-time based on this monitoring. The set-time corresponds to the onset of the coagulation phase in the milk curd based on a change in the evolution of the speckle contrast parameter. In other terms, the moment at which a change in the evolution of the speckle contrast parameter is observed corresponds to the onset of the coagulation phase, that is, to the set-time. It will be readily understood that subsequently to the determination of the set-time, the method 100 may involve any appropriate action based on this determination. In some embodiment, a display (not shown) may provide a graph similar to the one of FIG. 1 for viewing by the cheesemaker or other user. The graph may be built over time, or shown after the onset of the coagulation phase, in some cases, a time stamp marking the set-time may be provided to the cheese maker, for example as the corresponding time of day or as a counter marking the time having passed since the set-time.

In some embodiments, an additional step of determining 150 the suggested cuttime for the milk curd based on the evolution of the speckle contrast parameter may be provided. By way of example, the suggested cut-time for the milk curd based on the set-time and on at least one pre-selected condition may be performed. The pre-selected condition or conditions may for example include a predetermined delay between the set-time and the cut-time. In some variants, cutting may occur when the viscoelasticity has stabilized, nearly stabilized or reached a particular firmness value. In some embodiments, the monitoring of the viscoelasticity of the milk curd as explained above may be continued and the cuttime selected based on the observation of a further change in the evolution of the speckle contrast parameter.

In other embodiments the suggested cut-time may be determined directly from the evolution of the speckle contrast parameter without the intermediate determination of the set-time. In some variants, the monitoring of the speckle contrast and its time evolution (i.e. first time derivative, second time derivative) can be used to provide a ratio that is indicative of its value and its rate of change. This rate of change may be used as indicative of the cut-time having been reached. Referring back to FIG. 1 , is some embodiments, the suggested cut-time may be selected at a moment when the time evolution of the speckle contrast parameter flattens out, after the initial plateau of the set-time. Any suitable data processing technique allowing to analyse the behavior of the time evolution of the speckle contrast parameter may be used. The monitoring of the speckle contrast parameter is interrupted after the suggested cut time has been signaled or upon an intervention of an operator once the cheesemaker decides to proceed with the cutting. As seen in FIG. 1 , after the suggested cut-time, the speckle contrast parameter remains stable, such that the representation of its time evolution would show a relatively flat line if the monitoring was continued.

It will be readily understood that subsequently to the determination of the set-time and/or of the suggested cut-time, the method 100 may involve any appropriate action based on this determination. By way of example, after the suggested cuttime has been determined, the method may further comprise further an additional step of signaling an inception of the suggested set-time to an operator of the milk product manufacturing process. The signaling may may auditory and/or visual, and may for example involve an audible alarm, a visual alarm, an email, a text message, a pager call, a phone call, an activation of a portable vibrating system, and the like.

Different implementations of the steps of obtaining 106 the scattering data sets and calculating 108 and monitoring 110 the speckle contrast parameter may be envisioned, examples of which are provided below. In some implementations, the calculation of the speckle contrast parameter may be achieved based on one of two methods of analysis, a spatial analysis and temporal analysis. Both analyses can also be combined to achieve similar results.

Spatial analysis for calculating the speckle contrast parameter

Referring to FIG. 4, there is presented a flow chart of the monitoring 104 a viscoelasticity of the milk curd using a spatial analysis, according to some embodiments. In this variant, the monitoring 104 first includes acquiring 112, over time, a plurality of 2D images each representative of a scattering of the coherent light beam in the milk curd. Preferably, in such variants, the light sensing device may be embodied by a camera or other form of 2D matrix of pixels. By way of example, the camera can be an 8-bit camera or a 16-bit camera, that is, the camera can capture images with 256 levels of each colour in an 8-bit format. In some variants, the camera is configured to capture images with a pixel dimension of 250x250. Pixel dimensions express the number of pixels horizontally and vertically that define the resolution of the image. In other variants, a single mode optical fiber with a photodiode could be considered as a suitable 1 -pixel camera.

In some implementations, this step is realized by capturing several images at a fixed interval over a given acquisition duration. Each 2D image corresponds to a light scattering data set representing the distribution of the scattered light over an observation area of the milk curd covered by the field of view of the camera. The observation area may be selected so as to enable individual pixels to be resolved, while providing enough data for reliable statistical analysis. Each 2D image is preferably acquired over an exposure period or exposure time of the order of a millisecond or shorter. The exposure period, which can also be referred to as the exposure time or capture time, may be understood as the length of time over which photons scattered by the milk curd are sensed by the sensing device, as explained above. In typical implementations, the camera may have a shutter that opens for a period of about one millisecond each time it takes an image. By definition, in imaging applications, the amount of light that reaches each pixel of the camera corresponds to the exposure, and the accumulated light intensity on each pixel during one exposure period corresponds to a data point of the corresponding light scattering data set.

A plurality of 2D images, embodying a plurality of light scattering data sets, is obtained by repeating the image capture process explained above at an image capture frequency. The image capture frequency can vary according to the needs of the application; In some implementations, it is possible to capture an image every millisecond. In other variants, the capture frequency may be longer (e.g., 5ms, 10ms, every second, etc.). The capture frequency may be constant of may vary in time in some variants.

Each 2D image includes a speckle pattern. By way of example, FIGs. 5A, 5B and 50 show three examples of 2D images 40a, 40b and 40c acquired towards the beginning, the middle and the end of the pre-coagulation phase, respectively. Each 2D image 40a, 40b and 40c has a corresponding speckle pattern 42a, 42b and 42c. The contrast level of the speckle pattern is seen to increase over time.

Referring back to FIG. 3, in some embodiments, the optical system 20 may include a controller 29 operatively connected to the camera 26 (embodying a light detecting device) and configured to obtain therefrom, over time, the plurality of light scattering data sets. The controller 29 may be integral or external to the camera 26. The controller 29 sends control signals to the camera 26 to open and close the shutter and perform any task associated with the acquisition process. The optical system 20 further includes a processing unit 30 such as a server, computer, tablet, or the like. The processing unit 30 can be a single machine or a cluster of machines interconnected with each other, and the controller 29 may in some embodiments be integral to the processing unit 30. The processing unit 30 can include physical machine(s) and/or virtual machine(s). For instance, the camera 26 can be connected to a computer or computers positioned next to the milk curd tank and transmit the images it acquires to these computers. In other examples, the camera 26 can be wireless and transmit the images to virtual computers stored in a cloud infrastructure at another location. In one variant, the processing unit 30 may be a microcontroller in operative communication with the camera 26. The microcontroller can have one or more processor(s) and memory. The memory can store a program or code instructions to carry out the calculation of a speckle contrast parameter when executed by the processor(s).

Each image captured by the camera 26 can be stored by the processing unit 30 before being analyzed. In one variant, images are stored in a directory on a microcontroller or computer. When the program is run (i.e., the processor performing the calculation), it is configured to extract the images from this directory and perform the associated operations, as detailed below. In other cases, although not explicitly written down, images may be saved in a relational or non-relational database.

Referring to FIGs. 4, 6 and 6A, after acquiring 112 a plurality of 2D images, the monitoring 104 of the viscoelasticity of the milk curd involves calculating 114 the speckle contrast parameter of the speckle pattern in each 2D image. In some implementations, this calculation may be performed by dividing 116 the 2D image, or a selected zone of the 2D image into a plurality of tiles. In the illustrated example of FIG. 6A, the central zone of a 2D image 40a shown in FIG. 6 is shown divided into 25 tiles of equal size, but one skilled in the art will readily understand that the number of tiles and format could vary. By dividing into tiles, it is meant that the 2D image is segmented into equal portions such that the local mean intensity of the detected light is approximately constant, so that each portion can be analyzed. The tiles are added to a list and a speckle contrast value is then calculated 118 for each tile. In some embodiments, the speckle contrast value may be obtained by subtracting the minimum light intensity of the tile from the maximum intensity and dividing by the average intensity for the tile. A standard deviation can be calculated, normalized or not by the average intensity. An histogram can be analyzed in more details to extract more information than the average and standard deviation (kurtosis, etc..), as explained below. Other definitions of the speckle contrast value may be envisioned.

The method may then involves calculating 120 an image average of the speckle contrast values for the plurality of tiles. In some embodiments, this step may be repeated for each image until the size of the contrast data list is greater than or equal to a threshold value. For example, if the threshold value is 100 and 400 images per second are accumulated in the directory after the 50th image, the average contrast is measured. The calculated image average of the speckle contrast value, or a value directly derived therefrom, is outputted as the speckle contrast parameter and for example correspond to a mark on a graph similar to the one of FIG. 1 at the corresponding moment in time (ex : at 5 minutes the contrast is 0.45).

In some cases, it can be considered that once the speckle contrast parameter for a moment in time is outputted, the images that were previously analyzed are removed or deleted from the directory. For instance, once analyzed, these images can be moved to a different directory on a different device and replaced with new images captured by the camera. In some implementations, storing these images in another directory for future use may be useful for automation purposes. For example, these images can be used as a training dataset to train a suitable machine learning algorithm to determine the set-time and/or cut-time.

A more detailed example of carrying out the spatial analysis process is presented as follows (using Python syntax and terminology, where np=numpy):

- the images are extracted from the directory in which they have been stored and are then sorted by filename and placed in a list of images (named files);

- the processor loops through each image in the list and opens it;

- Once opened, the processor segments the image into tiles of the same size (show by the function tile mage) and add the tiles into a list of tiles;

- The processor then gets the tile at the centre of the list of tiles (img = tiles[C*C//2] where “//” means integer division without remainder);

- The time at which the image was captured is determined. If it's the first image, the time is set to 0.

- The image is then put in an array (a form of list) img_array = np.array(img) and then the mean of all the colours in the image is calculated (called first mean) firstMean = np.mean(np.mean(img_array)) (line 68). The first mean corresponds to the mean of the colours in the first image. Because it’s the first, it is stored and used as an initial value at time 0 - in other terms, it corresponds to the first image captured;

- The image array goes through a function image_contrast in which the contrast for each tile of the image is calculated ( np.std(tile) I np.mean(tile) , or can also be calculated simply with np.max(tile) - np.min(tile) I np.mean(tile)) and returned in a contrast list (maned contrasts).

- The processor determines the size of the contrast list is at least 100. For example, for 5 images you can have 125 tiles (25 tiles per image), which is equal to 125 contrast value for each of these tiles. If it’s more than 100, it returns the mean of all these contrasts and the time at which it was taken. This would result in a value such as at 5 minutes the contrast is 0.3.

Temporal analysis for calculating the speckle contrast parameter

Referring to FIG. 7, there is presented a flow chart of the monitoring 104 a viscoelasticity of the milk curd using a temporal analysis, according to some embodiments. In this variant, the monitoring 104 of the firmness of the milk curd first includes measuring 122, over an exposure period, a light intensity signal representative of a scattering of the coherent light beam in the milk curd. The exposure period is selected so as to be long enough to contain a variation in a speckle contribution to said light intensity signal. The method next includes calculating 124 a speckle contrast parameter from this light intensity signal. In some implementations, this may for example be carried out by calculating 126 a temporal contrast value over the corresponding exposure period, and outputting 128 this temporal contrast value as the speckle contrast parameter for the light intensity signal.

In typical implementations of the temporal analysis, the evolution of the speckle contrast parameter is evaluated for each pixel rather than over tiles. Typically, the temporal analysis involves looping through the images (light intensity signal values) from a given directory, determining whether the image is the first image taken (similar to what was done for the spatial analysis), and then calculating the contrast for the images. To do so, the numeral values of the pixels in each image are added into sequences of lists called stack_array (ex: values of the pixels of image 1 are added in a list with the values of the pixels of image 2 in the same list). The variance is calculated based on the standard deviation of the stack_array divided by the average (or mean) of the same stack_array. The variance is then flattened, meaning that the array (or list) is transformed into a single list or array (or a single dimension, since the calculation is based on a two-dimensional array). Finally, based on the variance, the mean standard deviation and skew are calculated along with the time of the time at which the initial images were taken.

In some implementations, it is possible to build a histogram from the data obtained through the temporal or spatial analysis. Typically, a histogram gives the intensity distribution across an image, i.e. , the number of pixels with each possible intensity level or across time at a given pixel. For example, in the case of an 8-bit grayscale image, as in this case, it is possible to have a total of around 256 possible intensities. The histogram can therefore be generated from the set of pixels in the collected images. For example, as explained in the previous code, one could stack the pixels of several images together and then calculate the histogram at that point in time. The processing unit could then save the histogram data in a directory or database and use it at another time (several seconds or minutes later) to compare it with a newly generated histogram, based on newly processed images representing a different state of the curd. By making this comparison, it would be possible to see the difference in pixel distribution and consequently the variation in speckles over time.

It should be noted that the algorithms described above may run continuously throughout the process of monitoring the firmness of the milk curd. In other words, in preferred embodiments the processing unit continuously receives images, processes them and returns a new contrast/time value for display. The monitoring of the evolution of the speckle contrast parameter can then occur. At each given exposure period, or after a certain number of images have been taken at a given interval, the algorithm (or the processor) returns a time and contrast value representative of the average (mean) contrast for these images.

The speckle contrast parameter can thus be outputted to the user for consultation. In some cases, the output can simply be the raw value and time. In the present case, the module is configured to accumulate the values it has determined (the set of time/contrast values) and generate a graph. In some other cases, the processing unit can be in communication with a front-end module, or a display module adapted to display the given value. For instance, the processing unit can be connected to a screen to display a graph to the user (on an LCD screen, for example).

In some cases, the processing unit can further be configured to determine that the result it outputs corresponds to a preset critical time or a preset threshold. For example, the processing unit can be configured to determine that the resistance value corresponds to or is proximal to a curing time and/or a cutting time. In this case, the processing unit can transmit an alert indicating the cutting or curing time. An alert can be transmitted in the form of a signal to a light (LED light), an audio queue, an e-mail alert on an employee's device, and so on.

Of course, numerous modifications could be made to the embodiments above without departing from the scope of protection.

Claims

Claims:

1 . A method for determining a suggested cut-time for a milk curd in a milk product manufacturing process, the method comprising: a. projecting a coherent light beam on the milk curd; b. monitoring a viscoelasticity of the milk curd, said monitoring comprising: i. obtaining, over time, a plurality of light scattering data sets each representative of a scattering of the coherent light beam in the milk curd over an exposure period; ii. in real time, calculating a speckle contrast parameter from the light scattering data set; and iii. monitoring the evolution of the speckle contrast parameter over time; and c. determining the suggested cut-time for the milk curd based on the evolution of the speckle contrast parameter.

2. The method according to claim 1 , wherein the milk product is a cheese product.

3. The method according to claim 1 or 2, wherein the plurality of light scattering data sets comprises a plurality of 2D images, the exposure period for each 2D image having a span of the order of a millisecond or shorter, each 2D image comprising a speckle pattern.

4. The method according to claim 3, wherein calculating the speckle contrast parameter for each 2D image comprises: a. dividing the 2D image into a plurality of tiles; b. for each tile, calculating a speckle contrast value; and c. calculating an image average of the speckle contrast values for the plurality of tiles of said 2D image, and outputting said image average as the speckle contrast parameter for said image.

5. The method according to claim 1 or 2, wherein obtaining a plurality of light scattering data sets comprises measuring, at different time intervals, a plurality of light intensity signals, the exposure period for each light intensity signal having a length sufficient to contain a variation in a speckle contribution.

6. The method according to claim 5, wherein calculating the speckle contrast parameter for each light intensity signal may comprise calculating a temporal contrast value over the corresponding exposure period, and outputting said temporal contrast value as the speckle contrast parameter for said light intensity signal.

7. The method according to any one of claims 1 to 6, wherein determining the suggested cut-time for the milk curd comprises: a. identifying a set-time corresponding to an onset of a coagulation phase in said milk curd based on a change in the evolution of the speckle contrast parameter; and b. determining the suggested cut-time for the milk curd based on the settime and on at least one pre-selected condition.

8. The method according to claims 7, wherein the at least one pre-selected condition on which the determining of the suggested cut-time is based comprises a predetermined delay between the set-time and the suggested cuttime.

9. The method according to claim 6 or 7, further comprising continuing the monitoring of step b. past the set-time, and wherein the at least one preselected condition on which the determining of the suggested cut-time is based comprises an observation of a further change in the evolution of the speckle contrast parameter after set set-time.

10. The method according to any one of claims 1 to 9, further comprising an additional step of signaling an inception of the suggested set-time to an operator of the milk product manufacturing process.

11 . A method for determining the onset of a coagulation phase in a milk curd in a milk product manufacturing process, comprising: a. projecting a coherent light beam on the milk curd; and b. during a pre-coagulation phase of said milk product manufacturing process, monitoring a viscoelasticity of the milk curd, said monitoring comprising: i. obtaining, over time, a plurality of light scattering data sets, each representative of a scattering of the coherent light beam in the milk curd over an exposure period; ii. in real time, calculating a speckle contrast parameter from each light scattering data set; and iii. monitoring the evolution of the speckle contrast parameter over time; whereby a moment at which a change in the evolution of the speckle contrast parameter is observed corresponds to the onset of the coagulation phase.

12. The method according to claim 11 , wherein the plurality of light scattering data sets comprises a plurality of 2D images, the exposure period for each 2D image having a span of the order of a millisecond or shorter, each 2D image comprising a speckle pattern.

13. The method according to claim 12, wherein calculating the speckle contrast parameter for each 2D image comprises: a. dividing the 2D image into a plurality of tiles; b. for each tile, calculating a speckle contrast value; and c. calculating an image average of the speckle contrast values for the plurality of tiles of said 2D image, and outputting said image average as the speckle contrast parameter for said image.

14. The method according to claim 11 , wherein obtaining a plurality of light scattering data sets comprises measuring, at different time intervals, a plurality of light intensity signals, the exposure period for each light intensity signal having a length sufficient to contain a variation in a speckle contribution.

15. The method according to claim 14, wherein calculating the speckle contrast parameter for each light intensity signal comprises calculating a temporal contrast value over the corresponding exposure period, and outputting said temporal contrast value as the speckle contrast parameter for said light intensity signal.

16. The method according to any one of claims 11 to 15, wherein said milk product is a cheese product.

17. An optical system for monitoring a viscoelasticity of a milk curd in a milk product manufacturing process, comprising:

- a light sensing device configured to detect the scattered light beam;

18. The optical system according to claim 17, wherein the coherent light source comprises a laser diode.

19. The optical system according to claim 17 or 18, wherein the light sensing device comprises a 2D matrix of pixels.

20. The optical system according to claim 17 or 18, wherein the light sensing device comprises a photodiode.