Cloth acceptance value regulation and control method

Abstract

The invention discloses a cloth acceptance value regulation and control method. The cloth acceptance value regulation and control method comprises the following steps: comparing a first characteristic standard value of a standard cloth image with a first characteristic to-be-detected value of a first to-be-detected cloth image; when a first difference between the first characteristic standard value and the first characteristic to-be-detected value is smaller than or equal to a first acceptance difference, classifying the first to-be-detected cloth image into a first category; when the first difference is greater than the first acceptance difference, judging whether the first to-be-detected cloth image belongs to a first category or not through a neural network model; when the neural network model judges that the first to-be-detected cloth image belongs to a first category, the first acceptance difference is adjusted, and the first category is corresponding qualified cloth. In this way, the situation that qualified cloth is classified as unqualified cloth can be reduced, and the calculation amount can be reduced.

Description

Technical Field

[0001] The present disclosure relates to a fabric acceptance value control technology, and more particularly to a fabric acceptance value control method. Background Art

[0002] Manual inspection of fabric defects has a low detection rate, takes a long time, and has high labor costs. In addition, the detection rate varies from person to person, there is no consistency standard, and it cannot be verified. Therefore, developing a technology that can overcome the above problems is an important issue in this field. Summary of the invention

[0003] The disclosed embodiment includes a method for controlling the acceptable value of fabric. The method includes: comparing a first characteristic standard value of a standard fabric image and a first characteristic value to be measured of a first fabric image to be measured; when a first difference between the first characteristic standard value and the first characteristic value to be measured is less than or equal to a first acceptable difference, classifying the first fabric image to be measured into a first category; when the first difference is greater than the first acceptable difference, determining whether the first fabric image to be measured belongs to the first category through a neural network model; and when the neural network model determines that the first fabric image to be measured belongs to the first category, adjusting the first acceptable difference, wherein the first category corresponds to qualified fabric.

[0004] In some embodiments, the fabric acceptance value control method further includes: when the neural network model determines that the first fabric image to be tested belongs to a second category different from the first category, maintaining the first acceptance difference, and judging whether a second fabric image to be tested belongs to the first category based on the maintained first acceptance difference, wherein the second category corresponds to unqualified fabric, and the second fabric image to be tested is different from the first fabric image to be tested.

[0005] In some embodiments, the fabric acceptance value control method further comprises: after adjusting the first acceptance difference, determining whether the second fabric image to be tested belongs to the first category according to the adjusted first acceptance difference.

[0006] In some embodiments, adjusting the first acceptable variance includes: increasing the first acceptable variance to the first variance.

[0007] In some embodiments, the first feature value to be measured corresponds to one of a horizontal length, a vertical length, at least one coordinate of at least one certain position, an average pixel value of a plurality of pixels in a block, and a pattern of the first fabric image to be measured.

[0008] In some embodiments, the fabric acceptance value control method further includes: comparing a second feature standard value of the standard fabric image and a second feature value to be measured of the first fabric image to be measured; when a second difference between the second feature standard value and the second feature value to be measured is less than or equal to a second acceptance difference, classifying the first fabric image to be measured into the first category; when the second difference is greater than the second acceptance difference, determining whether the first fabric image to be measured belongs to the first category through the neural network model; and when the neural network model determines that the first fabric image to be measured belongs to the first category, adjusting the second acceptance difference.

[0009] In some embodiments, adjusting the first acceptance variance includes increasing the first acceptance variance to the first variance, and adjusting the second acceptance variance includes increasing the second acceptance variance to the second variance.

[0010] In some embodiments, the acceptable value control method further includes: after adjusting the first acceptable difference and adjusting the second acceptable difference, determining whether a second fabric image to be tested belongs to the first category based on the adjusted first acceptable difference and the adjusted second acceptable difference, wherein the second fabric image to be tested is different from the first fabric image to be tested.

[0011] In some embodiments, the fabric acceptance value control method further includes: when the neural network model determines that the first fabric image to be tested belongs to a second category different from the first category, maintaining the second acceptance difference, and determining whether a second fabric image to be tested belongs to the first category based on the maintained second acceptance difference.

[0012] In some embodiments, the first feature value to be measured corresponds to one of a horizontal length of the first fabric image to be measured, a vertical length, at least one coordinate of at least one certain position, an average pixel value of multiple pixels in a block, and a pattern, and the second feature value to be measured corresponds to the horizontal length of the first fabric image to be measured, the vertical length, the at least one coordinate of the at least one certain position, the average pixel value of the multiple pixels in the block, and the other of the pattern. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 is a schematic diagram of a fabric acceptance value control system according to one embodiment of the present disclosure;

[0014] Figure 2 is a schematic diagram of a standard fabric image and a first fabric image to be tested according to an embodiment of the present disclosure;

[0015] Figure 3 is a flow chart of a method for controlling a fabric acceptance value according to one embodiment of the present disclosure;

[0016] Figure 4 A flowchart showing further details of a method for controlling fabric acceptance value according to one embodiment of the present disclosure;

[0017] Figure 5 A flowchart showing further details of a method for controlling fabric acceptance value according to one embodiment of the present disclosure;

[0018] Figure 6 The flowchart is a further detailed diagram of a method for controlling fabric acceptance value according to one embodiment of the present disclosure.

[0019]

Explanation of symbols

[0020] 100: Fabric acceptance value control system

[0021] 110: Processing device

[0022] 120: Camera device

[0023] 130: Transport device

[0024] FBS: Standard fabric

[0025] MS: Standard fabric image

[0026] FB1: The first fabric to be tested

[0027] M1: The first fabric image to be tested

[0028] FB2: The second fabric to be tested

[0029] M2: The second fabric image to be tested

[0030] MD: Neural Network Model

[0031] LS2, L12: horizontal length

[0032] LS1, L11: vertical length

[0033] PS1~PS3、P11~P13: Positioning points

[0034] BS3, B13: Block

[0035] DS, D1: Pattern

[0036] 300: Fabric Acceptable Value Control Method

[0037] OP31~OP36、OP41~OP42、OP51~OP56、OP61~OP62: Operation DETAILED DESCRIPTION

[0038] In this document, although the terms "first", "second", etc. are used to describe different components, the terms are only used to distinguish components or operations described by the same technical terms. Unless the context clearly indicates otherwise, the terms do not specifically refer to or imply an order or sequence, nor are they used to limit the present disclosure.

[0039] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present disclosure belongs. It will be further understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology and the present disclosure, and will not be interpreted as an idealized or overly formal meaning unless explicitly defined as such herein.

[0040] The terms used herein are only for the purpose of describing specific embodiments and are not restrictive. As used herein, unless the content clearly indicates, the singular forms "one", "an" and "the" are intended to include plural forms, including "at least one". "Or" means "and / or". As used herein, the term "and / or" includes any and all combinations of one or more related listed items. It should also be understood that when used in this specification, the terms "include" and / or "comprise" specify the presence and / or parts of the features, regions, entireties, steps, operations, elements, but do not exclude the presence or addition of one or more other features, regions, entireties, steps, operations, elements, parts and / or their combinations.

[0041] The following will disclose multiple embodiments of the present disclosure with the accompanying drawings. For the purpose of clear description, many practical details will be described together in the following description. However, it should be understood that these practical details should not be used to limit the present disclosure. In other words, in some embodiments of the present disclosure, these practical details are not necessary. In addition, in order to simplify the drawings, some conventional structures and elements will be depicted in a simple schematic manner in the drawings.

[0042] Figure 1 FIG. 1 is a schematic diagram of a fabric acceptance value control system 100 according to an embodiment of the present disclosure. In some embodiments, the fabric acceptance value control system 100 includes a processing device 110 , a camera device 120 , and a transport device 130 .

[0043] In some embodiments, the camera device 120 is used to photograph the standard fabric FBS to generate the standard fabric image MS, photograph the first fabric to be tested FB1 to generate the first fabric to be tested image M1, and photograph the second fabric to be tested FB2 to generate the second fabric to be tested image M2.

[0044] In some embodiments, the processing device 110 is used to determine whether each of the first test fabric FB1 and the second test fabric FB2 belongs to the first category or the second category according to the standard fabric image MS, the first test fabric image M1 and the second test fabric image M2. The first category corresponds to qualified fabric, and the second category corresponds to unqualified fabric.

[0045] In some embodiments, the processing device 110 is further configured to determine whether each of the first test fabric FB1 and the second test fabric FB2 belongs to the first category or the second category through the neural network model MD.

[0046] In some embodiments, the transport device 130 is used to transport the first fabric to be tested FB1 and the second fabric to be tested FB2 to corresponding positions according to their categories. For example, when the first fabric to be tested FB1 belongs to the first category, the transport device 130 transports the first fabric to be tested FB1 to the first position. When the first fabric to be tested FB1 belongs to the second category, the transport device 130 transports the first fabric to be tested FB1 to a second position different from the first position.

[0047] Figure 2 FIG. 1 is a schematic diagram of a standard fabric image MS and a first fabric image M1 according to an embodiment of the present disclosure. In some embodiments, the processing device 110 generates a first characteristic standard value according to the standard fabric image MS, and generates a first characteristic value according to the first fabric image M1.

[0048] like Figure 2 As shown, the first characteristic standard value may correspond to one of the horizontal length LS2 and the vertical length LS1 of the standard fabric image MS, the coordinates of the positioning points PS1-PS3, the average pixel value of a plurality of pixels in the block BS3, and the pattern DS. The first characteristic to-be-measured value may correspond to one of the horizontal length L12 and the vertical length L11 of the first to-be-measured fabric image M1, the coordinates of the positioning points P11-P13, the average pixel value of a plurality of pixels in the block B13, and the pattern D1.

[0049] In some embodiments, the processing device 110 is used to perform calculations based on the features of the standard fabric image MS to generate positioning points PS1-PS3 of the standard fabric image MS, and to perform calculations based on the features of the first fabric image M1 to generate positioning points P11-P13 of the first fabric image M1 to be tested.

[0050] In some embodiments, the processing device 110 may determine the first feature value to be measured based on the distance between the positioning points P11-P13 and the boundary of the first fabric image M1 to be measured. The processing device 110 may also calculate the first difference between the first feature standard value and the first feature value to be measured based on the Mahalanobis distance between the patterns DS and D1.

[0051] Figure 3 FIG. 3 is a flow chart of a method 300 for controlling fabric acceptance value according to an embodiment of the present disclosure. Figure 3 As shown, the fabric acceptance value control method 300 includes operations OP31 to OP36. In some embodiments, the fabric acceptance value control method 300 can be Figure 1 The acceptable value control system 100 is shown to be implemented, but the disclosure is not limited thereto.

[0052] In operation OP31 , the first characteristic standard value of the standard fabric image MS and the first characteristic to be measured value of the first fabric image to be measured M1 are compared.

[0053] In operation OP32, it is determined whether the first difference between the first characteristic standard value and the first characteristic to be measured value is greater than the first acceptable difference. When the first difference is less than or equal to the first acceptable difference, operation OP33 is performed after operation OP32. When the first difference is greater than the first acceptable difference, operation OP34 is performed after operation OP32.

[0054] In operation OP33, the first fabric image M1 to be tested is classified into a first category. The first category corresponds to qualified fabrics.

[0055] In operation OP34, the neural network model MD is used to determine whether the first fabric image M1 to be tested belongs to the first category. When the neural network model MD determines that the first fabric image M1 to be tested belongs to the first category, operation OP35 is performed after operation OP34. When the neural network model MD determines that the first fabric image M1 to be tested belongs to a second category different from the first category, operation OP36 is performed after operation OP34. The second category corresponds to unqualified fabric.

[0056] In operation OP35, the first acceptance variance is adjusted.

[0057] In operation OP36, the first acceptable difference is maintained, and whether the second fabric image M2 to be tested belongs to the first category is determined according to the maintained first acceptable difference. The second fabric image M2 to be tested is different from the first fabric image M1.

[0058] For example, when the first difference is 5% and the first acceptable difference is 3%, the neural network model MD will determine whether the first fabric image M1 to be tested belongs to the first category. When the neural network model MD determines that the first fabric image M1 to be tested belongs to the second category, the first acceptable difference will be maintained at 5%, and the processor 110 determines the second fabric image M2 to be tested based on the first acceptable difference maintained at 5%. On the other hand, when the neural network model MD determines that the first fabric image M1 to be tested belongs to the first category, the processor 110 adjusts the first acceptable difference.

[0059] In some practices, when checking whether the fabric to be tested is qualified, it is necessary to overlap the image of the fabric to be tested with the image of the standard fabric and then manually compare them. The above-mentioned procedures are cumbersome and the labor cost is high. In addition, when the inspection personnel are inexperienced, they may accidentally classify qualified fabrics into unqualified categories.

[0060] Compared to the above approach, in the embodiment of the present invention, the first acceptable difference is first used to distinguish the first fabric image M1 to be tested. In the case that the first fabric image M1 to be tested does not meet the first acceptable difference, the neural network model MD is further used to determine whether the first fabric image M1 to be tested belongs to the first category, and the first acceptable difference is adjusted accordingly. In this way, the processor 110 can automatically adjust the first acceptable difference to reduce the situation where qualified fabrics are classified as unqualified. In addition, before using the neural network model MD, using the first acceptable difference for classification in operation OP32 can also reduce the amount of calculation of the neural network model MD.

[0061] Figure 4 FIG. 3 is a flow chart showing further details of a method 300 for controlling fabric acceptance value according to an embodiment of the present disclosure. Figure 4 As shown, the fabric acceptance value control method 300 further includes an operation OP42 performed after the operation OP35, and the operation OP35 may include the operation OP41.

[0062] In operation OP41, the first acceptable difference is added to the first difference.

[0063] In operation OP42, it is determined whether the second fabric image M2 belongs to the first category according to the adjusted first acceptable difference.

[0064] For example, when the first difference is 5% and the first acceptable difference is 3%, the neural network model MD will determine whether the first fabric image M1 to be tested belongs to the first category. When the neural network model MD determines that the first fabric image M1 to be tested belongs to the first category, the first acceptable difference will be adjusted to 5%, and the processor 110 determines the second fabric image M2 to be tested based on the first acceptable difference adjusted to 5%.

[0065] Figure 5 FIG. 3 is a flow chart showing further details of a method 300 for controlling fabric acceptance value according to an embodiment of the present disclosure. Figure 5 As shown, the fabric acceptance value control method 300 further includes operations OP51 to OP55 performed after the operation OP31.

[0066] In operation OP51 , the second characteristic standard value of the standard fabric image MS and the second characteristic to be measured value of the first fabric image to be measured M1 are compared.

[0067] In operation OP52, it is determined whether the second difference between the second characteristic standard value and the second characteristic to be measured value is greater than the second acceptable difference. When the second difference is less than or equal to the second acceptable difference, operation OP53 is performed after operation OP52. When the second difference is greater than the second acceptable difference, operation OP54 is performed after operation OP52.

[0068] In operation OP53 , the first fabric image M1 to be tested is classified into a first category.

[0069] In operation OP54, the neural network model MD is used to determine whether the first fabric image M1 to be measured belongs to the first category. When the neural network model MD determines that the first fabric image M1 to be measured belongs to the first category, operation OP55 is performed after operation OP54. When the neural network model MD determines that the first fabric image M1 to be measured belongs to the second category, operation OP56 is performed after operation OP54.

[0070] In operation OP55, the second acceptance variance is adjusted.

[0071] In operation OP56 , the second acceptable difference is maintained, and it is determined whether the second fabric image M2 to be tested belongs to the first category according to the maintained second acceptable difference.

[0072] Please refer to Figure 5 and Figure 2 In some embodiments, the first feature value to be measured corresponds to the horizontal length L12 of the first fabric image M1 to be measured, the vertical length L11, the coordinates of the positioning points P11~P13, the average pixel value of multiple pixels in the block B13, and one of the patterns D1, and the second feature value to be measured corresponds to the horizontal length L12 of the first fabric image M1 to be measured, the vertical length L11, the coordinates of the positioning points P11~P13, the average pixel value of multiple pixels in the block B13, and the other of the patterns D1.

[0073] Figure 6 FIG. 3 is a flow chart showing further details of a method 300 for controlling fabric acceptance value according to an embodiment of the present disclosure. Figure 6As shown, the fabric acceptance value control method 300 further includes an operation OP62 performed after the operation OP55 . The operation OP55 may include the operation OP61 , and the operation OP55 may be performed after the operation OP35 .

[0074] At operation OP61, the second acceptance difference is added to the second difference.

[0075] In operation OP62 , after adjusting the first acceptable difference and adjusting the second acceptable difference, it is determined whether the second fabric image M2 belongs to the first category according to the adjusted first acceptable difference and the adjusted second acceptable difference.

[0076] For example, when the first difference and the second difference are 5% and 6% respectively, each of the first acceptable difference and the second acceptable difference is 3%, and the neural network model MD determines that the first fabric image M1 to be tested belongs to the first category, the processor 110 adjusts the first acceptable difference and the second acceptable difference to 5% and 6% respectively. Subsequently, the processor 110 makes a judgment on the second fabric image M2 to be tested based on the first acceptable difference adjusted to 5% and the second acceptable difference adjusted to 6%.

[0077] In summary, the fabric acceptance value control method 300 can adjust various acceptance differences (such as the first acceptance difference and / or the second acceptance difference) according to the classification result of the first fabric image M1 to be tested, and determine whether the second fabric image M2 to be tested is qualified by the adjusted acceptance difference. Compared with the traditional method, the present disclosure can reduce the situation where the second fabric image M2 to be tested is judged as unqualified.

[0078] Although the present disclosure has been disclosed as above in the form of implementation methods, it is not intended to limit the present disclosure. Anyone with ordinary knowledge in the technical field may make slight changes and modifications without departing from the spirit and scope of the present disclosure. Therefore, the scope of protection of the present disclosure shall be determined by the definition of the attached claims.

Claims

1. A method for controlling the acceptable value of fabric, characterized in that: Include: Comparing a first characteristic standard value of a standard fabric image with a first characteristic to be measured value of a first fabric image to be measured; When a first difference between the first characteristic standard value and the first characteristic to be measured value is less than or equal to a first acceptable difference, classifying the first to-be-measured fabric image into a first category; When the first difference is greater than the first acceptable difference, determining whether the first fabric image to be tested belongs to the first category through a neural network model; and When the neural network model determines that the first fabric image to be tested belongs to the first category, the first acceptable difference is adjusted. The first category corresponds to qualified fabrics.

2. The method for controlling the fabric acceptance value according to claim 1, characterized in that: Also includes: When the neural network model determines that the first fabric image to be tested belongs to a second category different from the first category, the first acceptable difference is maintained, and a second fabric image to be tested is determined whether it belongs to the first category according to the maintained first acceptable difference. The second category corresponds to unqualified fabric, and the second fabric image to be tested is different from the first fabric image to be tested.

3. The method for controlling the fabric acceptance value according to claim 2, characterized in that: Also includes: After adjusting the first acceptable difference, determining whether the second fabric image to be tested belongs to the first category is performed according to the adjusted first acceptable difference.

4. The method for controlling the fabric acceptance value according to claim 1, characterized in that: Adjusting the first acceptable difference includes: The first acceptable variance is added to the first variance.

5. The method for controlling the fabric acceptance value according to claim 1, characterized in that: The first feature value to be measured corresponds to one of a horizontal length, a vertical length, at least one coordinate of at least a certain position, an average pixel value of a plurality of pixels in a block, and a pattern of the first fabric image to be measured.

6. The method for controlling the fabric acceptance value according to claim 1, characterized in that: Also includes: Comparing a second characteristic standard value of the standard fabric image with a second characteristic to be measured value of the first fabric image to be measured; When a second difference between the second characteristic standard value and the second characteristic to be measured value is less than or equal to a second acceptable difference, classifying the first to-be-measured fabric image into the first category; When the second difference is greater than the second acceptable difference, determining whether the first fabric image to be tested belongs to the first category through the neural network model; and When the neural network model determines that the first fabric image to be tested belongs to the first category, the second acceptable difference is adjusted.

7. The method for controlling the fabric acceptance value according to claim 6, characterized in that: in Adjusting the first acceptance variance includes adding the first acceptance variance to the first variance, and Adjusting the second acceptable variance includes adding the second acceptable variance to the second variance.

8. The method for controlling the fabric acceptance value according to claim 7, characterized in that: Also includes: After adjusting the first acceptable difference and adjusting the second acceptable difference, determining whether a second fabric image to be tested belongs to the first category according to the adjusted first acceptable difference and the adjusted second acceptable difference, The second fabric image to be tested is different from the first fabric image to be tested.

9. The method for controlling the fabric acceptance value according to claim 6, characterized in that: Also includes: When the neural network model determines that the first fabric image to be tested belongs to a second category different from the first category, the second acceptable difference is maintained, and a second fabric image to be tested is determined whether it belongs to the first category based on the maintained second acceptable difference.

10. The method for controlling the fabric acceptance value according to claim 6, characterized in that: in The first feature value to be measured corresponds to one of a horizontal length, a vertical length, at least one coordinate of at least one certain position, an average pixel value of a plurality of pixels in a block, and a pattern of the first fabric image to be measured, and The second feature value to be measured corresponds to another one of the horizontal length of the first measured fabric image, the vertical length, the at least one coordinate of the at least one fixed point, the average pixel value of the multiple pixels of the block, and the pattern.

Images