EP1678370A1

EP1678370A1 - Method for processing signals obtained by scanning planar textile structures

Info

Publication number: EP1678370A1
Application number: EP04761953A
Authority: EP
Inventors: Ian George; Sandra Edalat-Pour; Karl-Ludwig Schinner
Original assignee: Uster Technologies AG
Current assignee: Uster Technologies AG
Priority date: 2003-10-16
Filing date: 2004-10-07
Publication date: 2006-07-12
Also published as: JP2007522349A; WO2005035862A1; CN1867727A

Abstract

The invention relates to a method for processing signals that are obtained by scanning planar textile structures (1). The aim of the invention is to create a method that allows for a very differentiated evaluation of the defects (4, 5, 6) in a planar structure while leading to specific actions based on the identified defects. Said aim is achieved by first determining value ranges for parameter values, which define defect categories in the planar structure, whereupon the distribution of the defects in the planar structure is determined for defect categories in the planar structure, and an action is taken in relation to the planar structure according to the determined category and the distribution of the defects in the planar structure.

Description

Process for processing signals obtained by scanning textile fabrics

The invention relates to a method for processing signals which are obtained by scanning textile fabrics.

From EP 1100989, for example, a method for assessing defects in textile fabrics is known, with which the admissibility of defects in fabrics is determined on the basis of the length and the contrast of the defects compared to the defect-free fabric. The longer the error and the greater the contrast, the more likely it is that the error in question is undesirable. Basically, the same actions are then provided for each error that is recognized as such and undesirable, that is to say unacceptable. This means that an undesirable defect in a textile fabric means that the defect is removed if possible, or that the part of the fabric concerned is sold at a lower price or not at all and is therefore considered to be rejects.

A disadvantage of this known method can be seen, for example, in the fact that a clear boundary is drawn between permissible and impermissible errors. This limit is chosen based on the weighting of opposing influences such as economy and quality of the fabric. To ensure profitability, the limit should be chosen so that as few errors as possible lead to rejects. To ensure quality, all defects should be recognized as such and removed as far as possible, or the fabric must be counted towards the committee. These opposing influences mean that the decision about permissible and impermissible errors is an undifferentiated and difficult to find compromise.

It is therefore an object of the present invention, as characterized in the claims, to provide a method which enables a very differentiated evaluation of the defects in a flat structure and leads to targeted actions based on the detected defects. The object is achieved in that values for defined parameters, such as contrast, intensity, length, direction, etc., are derived from signals that arise when scanning textile fabrics. Limit values are also specified for the parameter values, which are used to determine errors in the fabric. For the errors, ie the parameters that characterize them, value ranges are defined that define categories of errors in the fabric. For each category of the defects in the fabric, the distribution of the errors in the fabric is determined, and depending on the category and the distribution of the errors in the fabric, an action is taken on the fabric if necessary, such as counting the errors, stopping the drive for the fabric, triggering an alarm, ignoring or marking errors, etc.

The advantages achieved by the invention can be seen, in particular, in the fact that it can be used to increase both the economy of the manufacture of a flat structure and the quality of the flat structure, since the errors and their distribution lead to actions which depend very much on how disruptive the errors for the specific intended use of the fabric are real. This allows the assessment of the errors to be adjusted very finely and differentially to all possible circumstances.

The invention is explained in more detail below with reference to an exemplary embodiment and figures, FIGS. 1 and 2 each representing part of the method and FIG. 3 a device suitable for carrying out the method.

1 shows a coordinate system with three axes x, y and z, the x-axis and the y-axis relating to the dimensions of an area structure examined. For example, the x-axis can thus lie transversely to the fabric and the y-axis indicate the longitudinal direction of the fabric. The z-axis is assigned to one or more parameters such as intensity, contrast, color, etc. of the fabric. This means that length measurements along the x and y axes and values along the z axis for intensity, contrast, color, etc. can be applied. 100 denotes a field that lies above the xy plane, which spans the x and y axes together. This field 100 indicates the level of the values of a parameter. Thus, for the distance between the xy plane and the field 100, this distance indicates the value of the parameter in question. When scanning such a sheet by sensors, lines are formed. Here are some lines with n to n + 4 and so on. When viewed closely, these lines connect, for example, centers 101, 102 of Pixels 103, 104, which here lie on line 105, which means that lines 105 to 112 indicate values of pixels in a simplified representation. 113 denotes a series of values for line n + 4 or line 110. The special deflection of lines 106-111 shows that these represent a special feature in the form of a hill in field 100, which can be recognized by the measured values along lines 106-111. This peculiarity can also be referred to as error 114. In the case of a woven fabric, the lines n can run, for example, in the direction of the weft threads or the warp threads.

2 shows the series of values 113 already known from FIG. 1, which consists of a sequence of samples 115, 116 etc. These are plotted on a y-axis, along which, for example, values for a time or a path can be entered. Values of electrical quantities such as current, voltage, etc., which are derived from the measured intensity, contrast or color, can be plotted along the z axis. Values for various parameters can be derived from the series of values 113. As such, the length or duration 117 of a signal section 118 or parameters 119 derived from the sampled values, such as contrast, intensity, etc., which are proportional to the deflections of the series of values 113, are particularly conceivable. Limit values 120, 121 are also to be specified for parameters 117 and 119, which serve to determine an error in the fabric. Here, the limit value 120 relates to directly measured values for the parameter 119 and the limit value 121 relates, for example, to a parameter 117 derived from the value series 113, such as here the length or duration of the signal section 118 which is monitored by the predetermined limit value 121.

In Figure 3 you can see a part of a textile fabric 1, which is designed as a web and is moved in its longitudinal direction, such as during manufacture on a weaving machine or when rewinding from one roll to another during goods inspection, equipment, etc . the case is. A scanning device 2 is provided in front of or behind the fabric 1, which optically scans the continuous fabric 1 in a manner known per se, for example. For the fabric 1, a drive 3 is provided, which can consist, for example, of a driven roller or a pair of rollers. Various types of possible errors can be seen on the fabric 1, such as a swarm of errors 4, periodically occurring errors 5a, 5b, 5c, 5d and a surface error 6. Along the fabric, a length measuring device 7 is also arranged, which for example consists of an impeller a path encoder or consists of an optically operating device. The length measuring device 7 can also be integrated in the scanning device 2. Various memories e, 9, 11, a counter 10, computers 12, 13, 14, an input and output unit 15 and an actuator 16 are provided for processing analog or digital signals from the scanning device 2, which are connected to one another and to the other elements are connected via connections 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27 and 28.

The mode of operation of the invention is as follows: Before starting to scan the fabric 1, it is preferable to specify what should and should not be regarded as an error. For example, parameters must first be selected which are to characterize errors and limit values for the selected parameters must be specified, the exceeding of which indicate errors in the fabric. Furthermore, for values of the selected parameters that exceed limit values, value ranges must also be specified that define categories for the errors. Then information about the permissible distribution of errors and actions that should be triggered if the specifications are exceeded must also be specified. This can be done via the input / output unit 15 by manual input. Part of this can also be predefined, ie these values are already stored in the memories. While the sheet 1 is moved past the scanning device 2 by the drive 3, it is scanned, for example, in the form of cells corresponding to lines n, n + 1, n + 2, n + 3, n + 4, etc. (FIG. 1) , whereby a value series is created for each line, such as the value series 113 for the line n + 4. These series of values consist of samples such. B. the samples 115, 116, etc. (Fig. 2). These samples 115, 116, etc. have a deflection or a value that can be eaten, for example, as a current or voltage value. In the present environment, however, these scanning values 115, 116 etc. also represent physical parameters such as brightness, contrast and intensity. They depend on which device or which measuring principle is used when scanning the fabric 1. We call these variables, as well as variables derived from them, such as the length or duration of the signal section 118, parameters. Errors 4, 5, 6 (FIG. 3) in fabric 1 are identified by such parameters and their values. Sampling values 115, 116 etc. are thus supplied to the memory 8 via the connection 17. The memory 8 is, for example, a FIFO memory and, from the values obtained in series, it again composes an image of a part of the fabric 1, as shown in FIG. 1, in which the errors due to the values of their parameters also appear. Limit values and all information that can be used to assign errors to a specific error category belonging to a group of error categories, which will be described in more detail later, are stored in the memory 9. The computer 12 receives current parameter values via the connection 18 and via the connection 19 limit values at which the current parameter values are to be measured. By comparing the parameter values available to him with the predetermined limit values, he can assign each error to a category and, via the connection 21, provide the counter 10 with a corresponding indication which specifies the determined category. Here, the errors in each category are counted over a reference length of the fabric 1 specified by the user in order to determine whether there is a swarm. The scanned length is fed via connection 23 from length measuring device 7 to computers 12 and 13. The computer 13 calculates the number of errors per length, for example for each error category, and delivers this number to the computer 14 via the connection 24. This receives a specification from the memory 11 via the connection 25, which has stored a permissible number of errors calculated for each category, calculated per length of the fabric 1. This permissible number is fed to the memory 11 via a connection 29 from the input / output unit 15 and stored therein. From the current number from the connection 24 and the predetermined number from the connection 25, the computer 14 determines whether the actuator 16 should be instructed via the connection 26 to carry out an action or whether information is only sent to the input / output via the connection 28. Output unit should be submitted. The type of information or an action and how categories should be defined are described below. It is therefore possible according to the invention to use the input / output unit 15 to specify the type of errors or error categories which depend on the type of the textile fabric 1 present and are specified specifically. The following error categories are conceivable for a fabric, for example:

- short warp errors

- medium warp errors

- long warp defects

- short shot errors

- medium shot errors

- long shot errors

- gang fault

- edge defects

- Partial shot error

- small area errors

- medium area error

- large area errors

The user can also define error categories himself and enter corresponding value ranges for the parameters. For a knitted fabric, these could be the following error categories:

- holes

- Thin spots

- Thickening

- Stains

All of these errors have values for parameters that define the relevant category, such as:

- place

- dimension

- Average brightness, intensity or contrast, which are also input via the input / output unit 15 and assigned to the selected categories.

The parameters which are temporarily stored in the memory 8 are compared in the computer 12 with predetermined values for these parameters from the memory 9, which define categories, and thus assigned categories. For example, the length of an error in the weft direction could be specified as a parameter, with 0.5, 3, and 10 cm, for example, being the predefined values. If the measured length of the shot error is greater than 10 cm, it goes into the category "long shot error", if its length is between 3 and 10 cm, it goes into the category "medium shot error" and its length is between 0.5 and 3 cm, it falls into the category "short shot error". Are determined in addition to the above-mentioned categories of a first type, the ^~ by the characteristics of errors, even of another type categories should be specified or defined, which characterize the distribution of errors on the sheet. 1 As examples of this further type of category, we give: - individual errors that occur without a recognizable rule or errors that the user wants to have displayed from the first occurrence, - periodic errors that occur at regular intervals and - error swarms that arise there are localized clusters of errors consisting of individual errors that could be tolerable in isolation. For this purpose, a minimum number of errors should be entered via the input / output unit 15, for example for periodic errors, from which errors that occur regularly are to be considered periodic. For a swarm of errors, the number of errors should in turn be entered via the input / output unit 15. The number of errors here relates to a minimum number per reference length for the fabric 1, from which the errors can be regarded as a swarm of errors. It is possible to individually select the value ranges of the individual parameters for the definition of the categories and the distribution.

After the error categories of the first type, which are firmly entered in advance by the associated value ranges, and determined by suitable inputs, the device according to FIG. 1 determines the categories of the further type, which is determined by the determined distribution of the errors. The device again determines the actions to be carried out from both categories. For this purpose, the properties of the errors and the distribution of the errors are processed in the computer 14 in accordance with a predetermined program and an action is subsequently determined. As possible examples of actions we mention:

- Counting

- Trigger alarm

- Stop the drive

- To ignore

- To mark.

The computer 14 can transmit counted errors via the connection 28 to the input / output unit for display. This can also apply to an alarm. If the drive 3 is to be stopped, a corresponding signal is sent to the actuator 16 via the connection 26 and to the drive 3 via the connection 27.

The user can use the input / output unit 15 to enter values for the parameters which prescribe categories of the first type which appear to be essential for each conceivable fabric. However, it is also possible to enter parameters instead of directly by deriving the signals from the scanning device 2, e.g. Measured values are output to the input / output unit 15 and the memory 9 via the connection 20 ′ and these measured values are assigned to a category via the input / output unit 15.

The elements shown in the figure, such as memory, computer, etc., can be understood as function blocks of a data processing program. However, they can also be designed as individual fixed components of a circuit for signal processing. The method described above can also be shown in the table below. In this example, only one parameter is mentioned, which is defined by value ranges. However, other parameters could also be listed that are defined by other values or ranges. An error category is always defined by a combination of preferably several parameters with value ranges for each parameter.

Error category Parameter Action Value range Single swarm Periodic

Warp errors short 0.5 - 3 cm _! D medium 3 - 10 cm + + | long> 10 cm D D D

Short shot error 0.5 - 3 cm - +! medium 3 - 10 cm + +! long> 10 cm +! !

Partial shot shot error> 3cm + +! border on edge «I

Banding error ³ Λ fabric width DDD> 3 weft threads

Edge error weft direction + D D

Area errors small 0 <3 cm + D medium 0 3 - 10 cm + D large 0> 10 cm + D

Actions: + = count,! = Output alarm, D = switch off drive, - = ignore.

Claims

claims:

1. A method for processing signals which are obtained by scanning textile fabrics (1), characterized in that values for preselected parameters (117, 119) are derived from the signals, that limit values (120, 121 ) are specified, which are used for the determination of defects in the fabric, that for the values of the parameters value ranges are determined, which define categories of errors in the fabric, that for categories of errors in the fabric the distribution of the errors in the fabric is determined and that in Depending on the determined category and the distribution of errors in the fabric, an action in connection with the fabric is carried out.

2. The method according to claim 1, characterized in that an action from a group comprising counting the errors, neglecting the errors, switching off a drive for the fabric, marking the errors and triggering an alarm is selected as the action on the fabric becomes.

3. The method according to claim 1, characterized in that categories from a group containing warp defects, weft defects, surface defects, edge defects are determined as categories of defects in the fabric.

4. The method according to claim 1, characterized in that those are derived as parameters from a group containing at least length, width, contrast, intensity, diameter, direction, etc.