A Method for Evaluating Texture Consistency of Heather Gray Fabric Based on Random Sampling

Abstract

This invention relates to the field of fabric style and quality inspection technology, and in particular to a method for evaluating the texture consistency of heather gray fabric based on random sampling. The method includes the following steps: image acquisition and preprocessing; random image sampling; Gaussian filtering of the new image; calculation of texture structure features; calculation of texture consistency index; and texture consistency evaluation. This invention combines random image sampling, Gaussian filtering, and extraction of texture structure features to quantitatively calculate the texture consistency of heather gray fabric. It does not rely on human experience and achieves accurate and efficient evaluation of the texture consistency of heather gray fabric, exhibiting strong adaptability and versatility.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fabric style and quality detection, and particularly relates to a random sampling-based hemp gray fabric texture consistency evaluation method. BACKGROUND

[0002] Hemp gray yarn is generally a yarn composed of two or more different colored fibers or filaments. It can be directly knitted or woven into hemp gray style fabric, presenting a natural gray or mottled effect. Hemp gray fabric has been widely used in the field of clothing and home supplies due to its unique appearance and environmental performance. As a core measure of its quality, hemp gray fabric appearance consistency not only directly shapes the visual style of the fabric, but also deeply affects the overall quality of the end product. This indicator runs through the whole chain of fabric from design concept, production and manufacturing to consumer application, and is the key hub connecting the design concept, production process and consumer aesthetic needs. It is directly related to the acceptance of products in the market and brand value. Therefore, it is of great application value to study the evaluation method of hemp gray fabric appearance consistency.

[0003] For a long time, the evaluation of hemp gray fabric appearance consistency has been highly dependent on human experience, which is low in efficiency and strong in subjectivity. There are also a small number of existing technologies that do not rely on human experience to detect fabric quality. For example, a degree thesis (Ma Baolin. Research on quality detection of sliver based on image processing[D]. Wuxi: Jiangnan University, 2024) calculates the uniformity of the cotton web by using the coefficient of variation (a statistical measure used to measure the degree of dispersion of data, defined as the ratio of the standard deviation to the average value) in view of the difference in the light transmittance of the cotton web of the carding machine. However, this method is dependent on the image gray scale range and sensitive to noise, and cannot be directly applied to the evaluation of hemp gray fabric texture consistency.

[0004] In the current consumer market, which increasingly pursues high-quality and standardized products, it is of great significance to develop an accurate and efficient evaluation method for hemp gray fabric texture consistency. SUMMARY

[0005] The purpose of the present application is to solve the above-mentioned problems existing in the prior art, and to provide a random sampling-based hemp gray fabric texture consistency evaluation method.

[0006] To achieve the above-mentioned purpose, the technical solutions adopted by the present application are as follows:

[0007] A random sampling-based hemp gray fabric texture consistency (i.e. hemp gray effect consistency or uniformity) evaluation method, comprising the following steps:

[0008] (1) Image acquisition and preprocessing;

[0009] Obtain image A of the heather gray fabric, and perform Gaussian filtering on image A to obtain the filtered image A. g Gaussian filtering can effectively remove noise from images, smooth images, and highlight texture features in inconsistent areas.

[0010] (2) Random sampling of images;

[0011] Image C is obtained by performing N+1 random samplings on image A, where N≥10;

[0012] In each random sampling, image A is transformed into image B. In the first to Nth random samplings, image B is used as image A for the next random sampling after each random sampling. After the Nth random sampling, image B is used as image C.

[0013] (3) Gaussian filtering of the new image;

[0014] Applying a Gaussian filter to image C yields the filtered image C. g ;

[0015] (4) Calculation of texture structure features;

[0016] For image A g Perform a sliding window operation and calculate F for each sliding window. a F a The mean of all pixels within each sliding window and variance The ratio of F to all F a Calculate the average value to obtain ;

[0017] For image C g Perform a sliding window operation and calculate F for each sliding window. c F c The mean of all pixels within each sliding window and variance The ratio of F to all F c Calculate the average value to obtain ;

[0018] (5) Calculation of texture consistency index;

[0019] The formula for calculating the texture consistency index U is as follows:

[0020] ;

[0021] (6) Texture consistency evaluation;

[0022] Determine whether U≤0.4 is true. If it is true, the texture consistency is good; otherwise, the texture consistency is poor.

[0023] The principle of this invention is as follows:

[0024] When the texture consistency of the heather gray fabric image is poor, the heather gray fabric presents a mottled appearance, and this mottled distribution is disordered and of varying sizes. After randomly sampling such heather gray fabric, the obtained image will further highlight this uneven mottled appearance, making the texture consistency worse than before random sampling. After Gaussian filtering the images before and after random sampling, and then extracting their texture structure features Fa and Fc respectively, since the original image before random sampling already has poor texture consistency, random sampling will further amplify this difference, resulting in a larger U value.

[0025] If the image texture of the heather gray fabric has good consistency, the appearance of the image texture remains basically the same before and after random sampling, the difference between the extracted texture structure features is minimal, and the obtained U value will also be smaller.

[0026] As a preferred technical solution:

[0027] As described above, in the random sampling-based method for evaluating the texture consistency of heather gray fabric, in step (1), the size of image A is 256 pixels × 256 pixels.

[0028] As described above, in a random sampling-based method for evaluating the consistency of texture of heather fabric, in step (2), the process of converting image A into image B during a single random sampling is as follows: After dividing image A into K image blocks of size 8 pixels × 8 pixels in a non-overlapping manner, for each image block, a new image block of the same size as the image block is randomly extracted from image A, and the image block is replaced with the new image block to obtain image B.

[0029] As described above, in the method for evaluating the texture consistency of heather gray fabric based on random sampling, step (2) involves converting image A into image B during a single random sampling as follows:

[0030] (a) Divide image A into K non-overlapping image blocks of size 8 pixels × 8 pixels;

[0031] (b) Randomly generate two integers in the interval [1, M-8], denoted as rx and ry. Take the left vertex of image A as the origin of coordinates, extract a new image block with the left vertex coordinates (rx, ry) and a size of 8 pixels × 8 pixels from image A, and replace one of the original image blocks with it.

[0032] (c) Repeat step (b) until all the original image blocks are replaced by new image blocks, thus obtaining image B.

[0033] As described above, in the random sampling-based method for evaluating the texture consistency of heather gray fabric, the parameters of the Gaussian filter in steps (1) and (3) include: filter kernel size 5 and standard deviation 1.

[0034] As described above, in the random sampling-based method for evaluating the texture consistency of heather gray fabric, in step (4), the window size is 16 pixels × 16 pixels, the step size is 1 pixel, and the sliding method is: first from left to right, then from top to bottom.

[0035] Beneficial effects:

[0036] This invention first randomly scrambles the texture of the original image through multiple random samplings to obtain a scrambled image. Then, Gaussian filtering is used to process both the original image and the scrambled image to weaken texture features with good consistency while strengthening texture features in inconsistent areas. Next, the texture structure features of the original image and the scrambled image are extracted. Finally, the extracted texture structure features are used to quantitatively calculate the texture consistency of the heather gray fabric. The calculation process does not rely on human experience, achieving an accurate and efficient evaluation of the texture consistency of the heather gray fabric, and has strong adaptability and versatility. Attached Figure Description

[0037] Figure 1 It is a random sampling image and texture structure feature image of Example 1;

[0038] Figure 2 These are random sampling images and texture structure feature images from Example 2;

[0039] In the figure, part A is image A obtained by random sampling, and part C is image C obtained by random sampling; F a Part is F a The corresponding texture structure feature map, F c Part is F c The corresponding texture structure feature map. Detailed Implementation

[0040] The present invention will be further described below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0041] Example 1

[0042] A method for evaluating the texture consistency of heather gray fabric based on random sampling, the specific steps of which are as follows:

[0043] (1) Image acquisition and preprocessing;

[0044] Randomly select an image A of a heather gray fabric, image A is as follows. Figure 1 As shown in part A; Gaussian filtering is applied to image A to obtain the filtered image A. g Image A has dimensions of 256 pixels × 256 pixels, and the Gaussian filter parameters are a kernel size of 5 and a standard deviation of 1.

[0045] The manufacturer of the heather gray fabric is Jiangsu Hengli Chemical Fiber Co., Ltd., and the brand name is 23W079-1;

[0046] (2) Random sampling of images;

[0047] Image C is obtained by performing N+1 random samplings on image A, where N is 10. Image C is as follows: Figure 1 As shown in section C;

[0048] In each random sampling, image A is transformed into image B. In the first to Nth random samplings, image B is used as image A for the next random sampling after each random sampling. After the Nth random sampling, image B is used as image C.

[0049] The steps to transform image A into image B during a single random sampling are as follows:

[0050] (a) Divide image A into K non-overlapping image blocks of size 8 pixels × 8 pixels, where K is 1024;

[0051] (b) Randomly generate two integers in the interval [1, M-8], denoted as rx and ry, where M is 256. Using the left vertex of image A as the origin, extract a new image block with the left vertex coordinates (rx, ry) and a size of 8 pixels × 8 pixels from image A, and replace one of the original image blocks with it.

[0052] (c) Repeat step (b) until all the original image blocks are replaced by new image blocks, thus obtaining image B;

[0053] (3) Gaussian filtering of the new image;

[0054] Applying a Gaussian filter to image C yields the filtered image C. g Gaussian filter parameters: kernel size 5, standard deviation 1;

[0055] (4) Calculation of texture structure features;

[0056] For image A g Perform a sliding window operation and calculate F for each sliding window. a F a The mean of all pixels within each sliding window and variance The ratio of F to all F a Calculate the average value to obtain ;F a The corresponding texture structure features are as follows Figure 1 F a Partially shown;

[0057] For image C g Perform a sliding window operation and calculate F for each sliding window. c F c The mean of all pixels within each sliding window and variance The ratio of F to all F c Calculate the average value to obtain ;F c The corresponding texture structure features are as follows Figure 1 F c Partially shown;

[0058] When using the sliding window, the window size is 16 pixels × 16 pixels, the step size is 1 pixel, and the sliding method is: first from left to right, then from top to bottom;

[0059] (5) Calculation of texture consistency index;

[0060] The formula for calculating the texture consistency index U is as follows:

[0061] ;

[0062] (6) Texture consistency evaluation;

[0063] Determine whether U≤0.4 is true. If it is true, the texture consistency is good; otherwise, the texture consistency is poor.

[0064] The final U obtained in Example 1 was 0.42, indicating poor texture consistency.

[0065] Example 2

[0066] A method for evaluating the texture consistency of heather gray fabric based on random sampling is basically the same as in Example 1, except that:

[0067] The manufacturer of the heather gray fabric is Jiangsu Hengli Chemical Fiber Co., Ltd., brand number 23W079-4; N is 15; Image A is as follows. Figure 2 As shown in part A, image C is as follows Figure 2 As shown in part C, F a The corresponding texture structure features are as follows Figure 2 F a As shown in part, F c The corresponding texture structure features are as follows Figure 2 F c Partially shown.

[0068] Example 2 yielded a final U of 0.37, indicating good texture consistency.

[0069] Human eye assessment method:

[0070] To evaluate the accuracy of the evaluation results of Examples 1 and 2, the texture consistency of the fabric samples was further evaluated using the human eye. The specific steps of the human eye evaluation are as follows:

[0071] (I) Fabric sample preparation: Fabric sample A and fabric sample B, each 25cm long and 25cm wide, are flat and fixed on a non-reflective black background board to ensure uniform lighting.

[0072] (II) Observation conditions: The inspector is 50cm away from the fabric sample, with his line of sight perpendicular to the surface of the fabric sample. The inspector observes the entire fabric sample and local areas with his eyes.

[0073] (III) Comparison of texture consistency: The texture consistency is determined by whether the gray depth distribution of the heather texture on the surface of fabric sample A and fabric sample B is uniform. If the texture consistency of fabric sample A is better than that of fabric sample B, the score is 1; if the texture consistency of fabric sample A is worse than that of fabric sample B, the score is -1; if the texture consistency of fabric sample A is equal to that of fabric sample B, the score is 0.

[0074] (IV) Final result judgment: The sum of the scores of the three testers, S, is taken as the result for judgment. If S>0, it means that the texture consistency of fabric sample A is better than that of fabric sample B; if S=0, it means that the texture consistency of fabric sample A is basically the same as that of fabric sample B; if S<0, it means that the texture consistency of fabric sample A is worse than that of fabric sample B.

[0075] Human eye assessment results:

[0076] The heather gray fabric of Example 1 is designated as fabric sample A, and the heather gray fabric of Example 2 is designated as fabric sample B. The obtained human visual evaluation result S=-3, indicating that the texture consistency of the heather gray fabric of Example 1 is worse than that of the heather gray fabric of Example 2.

[0077] The heather gray fabric of Example 2 is designated as fabric sample A, and the heather gray fabric of Example 1 is designated as fabric sample B. The obtained human eye evaluation result S=3, indicating that the texture consistency of the heather gray fabric of Example 2 is better than that of the heather gray fabric of Example 1.

[0078] Therefore, the results of the human eye assessment are consistent with those of Examples 1-2, indicating that the assessment results of Examples 1-2 are accurate.

Claims

1. A random sampling-based evaluation method for the consistency of the texture of a linen fabric, characterized in that, The method comprises the following steps: (1) image acquisition and pretreatment; An image A of the linen ash fabric is acquired, and a Gaussian filtering process is performed on the image A to obtain a filtered image A g ; (2) image random sampling; N+1 times of random sampling are performed on the image A to obtain an image C, wherein N is greater than or equal to 10; In each time of random sampling, the image A is converted into an image B; in the first to Nth times of random sampling, the image B obtained after each time of random sampling is used as the image A for the next time of random sampling; and after the Nth time of random sampling, the image B is used as the image C; The step of converting the image A into the image B in each time of random sampling comprises the following steps: (a) dividing the image A into K image blocks with a size of 8 pixels by 8 pixels in a non-overlapping manner; (b) randomly generating two integers in the interval [1, M-8], denoted as rx and ry, taking the left top vertex of the image A as the coordinate origin, extracting a new image block with a size of 8 pixels by 8 pixels from the image A and replacing one of the original image blocks with the new image block; (c) repeating step (b) until all the original image blocks are replaced by new image blocks, and the image B is obtained; (3) Gaussian filtering processing of the new image; Gaussian filtering is performed on the image C to obtain a filtered image C g ; (4) texture structure feature calculation; For image A g Perform sliding window operation, calculate F a for each sliding window a , F a is the ratio of the mean value of all pixel points in each sliding window and variance , average all F a to get ; For image C g Perform a sliding window operation and calculate F for each sliding window. c F c The mean of all pixels within each sliding window and variance The ratio of F to all F c Calculate the average value to obtain ; (5) texture consistency index calculation; The formula for calculating the texture consistency index U is as follows: ； (6) texture consistency evaluation; It is determined whether U≤0.4 is true or not; if yes, the texture consistency is good; otherwise, the texture consistency is poor.

2. The method according to claim 1, wherein, In step (1), the size of the image A is 256 pixels by 256 pixels.

3. The method according to claim 1, wherein, In steps (1) and (3), the parameters of the Gaussian filtering include: a filter kernel size of 5 and a standard deviation of 1.

4. The method according to claim 1, wherein, In step (4), in the window sliding operation, the window size is 16 pixels by 16 pixels, the step length is 1 pixel, and the sliding manner is: first from left to right, and then from top to bottom.

Images