A visual recognition system for spinning quality

Abstract

The present invention relates to the field of visual recognition, and specifically discloses a spinning quality visual recognition system, comprising a processor, a spinning product image acquisition module and a spinning quality visual recognition module which are communicatively connected to the processor, and are used to solve the problem that the convexity and concavity of a local shadow area cannot be determined when the spinning product image is recognized; the present invention collects spinning product images through a high-resolution camera, performs grayscale threshold segmentation on the spinning product images, divides the spinning product images into a plurality of image area blocks, removes non-shadow areas, obtains shadow areas in the spinning product images, distinguishes shadow areas as convex areas or concave areas, and evaluates the quality of the spinning products by constructing a spinning product quality evaluation model, which can effectively solve the problem that the convexity and concavity of a local shadow area cannot be determined when the spinning product image is recognized, and helps to improve the accuracy and reliability of the evaluation of the quality of the spinning products.

Description

Technical Field

[0001] The present invention relates to the field of visual recognition, and more particularly to a visual recognition system for spinning quality. Background Art

[0002] The traditional detection method for the quality inspection of spinning products often uses the eyes of operators to identify defects, but this detection method has the problems of high missed detection rate, high work intensity and being greatly affected by the subjective factors of operators. These problems will lead to a decline in the quality of textiles, thereby affecting the tactile comfort of customers. Applying visual recognition to the defect identification of spinning products improves the accuracy and efficiency of defect identification and enhances the ability to distinguish the types of defects. At present, when visually identifying spinning products, it is found that some decorative areas of spinning products with decorations often need to be uniformly designed with convex parts, but due to the mechanical friction or uneven arrangement between fibers during the weaving process, the spinning products have different degrees of concavity and convexity, which makes the clarity of the image defective. When comparing the shadow areas with more shadows in the image, it is impossible to determine the concavity and convexity of the local shadow areas, which affects the comfort of customers and affects the aesthetics of the image and appearance of the spinning products. In order to solve the above problems, a technical solution is provided. Summary of the invention

[0003] In order to overcome the above-mentioned defects of the prior art, the present invention provides a spinning quality visual recognition system, which collects spinning product images through a high-resolution camera, divides the spinning product images into several image area blocks through grayscale threshold segmentation, and eliminates the non-shadow area to obtain the shadow area in the spinning product image, and distinguishes the shadow area as a convex area or a concave area. The quality of the spinning products is evaluated by constructing a spinning product quality evaluation model, which can effectively solve the problem that the convexity and concaveness of the local shadow area cannot be determined when identifying the spinning product image, and helps to improve the accuracy and reliability of the evaluation of the spinning quality, so as to solve the problems raised in the above-mentioned background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A spinning quality visual recognition system comprises a processor, a spinning product image acquisition module and a spinning quality visual recognition module which are communicatively connected to the processor, wherein the spinning product image acquisition module is used to collect spinning product images through a high-resolution camera; the spinning quality visual recognition module is used to visually recognize the spinning product images and perform quality assessment, the spinning quality visual recognition module comprises an image processing unit, a shadow area segmentation unit, an area detection unit and a spinning quality assessment unit, the spinning quality assessment unit is used to assess the quality of the spinning product by constructing a spinning product quality assessment model, and assess the quality of the spinning product by acquiring the pore size abnormality value of the spinning product, the total area of ​​the spinning product image of the concave area defect value and the total area of ​​G pore image blocks in the concave area, and importing the pore size abnormality value of the spinning product, the total area of ​​the spinning product image of the concave area defect value and the total area of ​​G pore image blocks in the concave area into the spinning product quality assessment model, wherein the formula of the spinning product quality assessment model is:

[0006]

[0007] Where: zl f is the quality index of spinning products, aq x is the defect value of the concave area, yf x is the abnormal value of the pore size of the spinning product, S z is the total area of ​​the spinning product image with defect value in the concave area, S g is the total area of ​​the G pore image blocks in the concave region.

[0008] As a further solution of the present invention, the image processing unit is used to remove noise, enhance contrast and grayscale the spinning product image;

[0009] The shadow area segmentation unit is used to segment the spinning product image into a plurality of image area blocks through grayscale threshold segmentation, and remove the non-shadow area to obtain the shadow area in the spinning product image;

[0010] The region detection unit is used to distinguish whether the shadow region is a convex region or a concave region;

[0011] The spinning quality evaluation unit is used to evaluate the quality of spinning products by constructing a spinning product quality evaluation model.

[0012] As a further solution of the present invention, the shadow area segmentation unit includes a non-shadow area recognition subunit, a non-shadow area segmentation subunit and a shadow area extraction subunit, wherein:

[0013] The non-shadow area recognition unit is used to recognize the non-shadow area in the spinning product image;

[0014] The non-shadow region segmentation subunit is used for segmenting the non-shadow region in the spinning product image;

[0015] The shadow area extraction unit is used to extract the remaining image area in the spinning product image, and the remaining image area is the shadow area.

[0016] As a further solution of the present invention, the non-shadow area segmentation subunit is used to segment the non-shadow area in the spinning product image, segment the spinning product image into a number of pore image blocks according to the specified pore size of spinning, eliminate the pore image blocks in the non-shadow area according to the spinning target grayscale value, and leave the pore image blocks in the shadow area.

[0017] As a further solution of the present invention, the spinning target grayscale value is obtained by calculating the average grayscale range of each pore image block, and the grayscale range of the pore image blocks is arranged in descending order as a range sequence {K h (i h ,j h ),…,K2(i2,j2),K1(i1,j1)}, where K1(i1,j1) is the pore image block with the smallest grayscale range value, K2(i2,j2) is the pore image block with the second smallest grayscale range value, and K h (i h ,j h ) is the pore image block with the largest grayscale range value; the 40% range sequence {K 0.4h (i 0.4h ,j 0.4h ),…,K2(i2,j2),K1(i1,j1)}, where K 0.4h (i 0.4h ,j 0.4h ) is a pore image block whose grayscale range value is 40% of the maximum value; the grayscale range value of the spinning target grayscale value of the remaining range sequence is compared with the spinning target grayscale value, and the pore image block whose grayscale range value is greater than the spinning target grayscale value is taken as a shadow area and marked.

[0018] As a further solution of the present invention, the region detection unit includes a concave-convex region discrimination unit, a convex region output unit and a concave region output unit; the concave-convex region discrimination unit is used to distinguish the concave-convexity of the shadow region by calculating the grayscale mean of the pore image block in the shadow region, comparing the grayscale value of the pore image block with the grayscale mean, and comparing the grayscale value of the pore image block with the grayscale mean in the following manner:

[0019] when When , the pore image block is a convex area;

[0020] when When , the pore image block is a concave area;

[0021] Among them, s Hq is the gray value of the qth pore image block, is the grayscale mean of the pore image block.

[0022] As a further solution of the present invention, the spinning quality evaluation unit includes a pore size anomaly analysis unit, a convex area statistical unit and a quality evaluation unit. The pore size anomaly analysis unit is used to extract the pore area in each pore image block in the spinning product image, arrange the pore area of ​​each pore image block from large to small according to the numerical value, and calculate the average pore area of ​​each pore image block. The pore area of ​​each pore image block and the average pore area of ​​each pore image block are introduced into the pore size anomaly analysis calculation model to calculate the pore size anomaly value of the spinning product, wherein the calculation formula of the anomaly analysis calculation model is:

[0023]

[0024] Where: yf x is the abnormal value of the pore size of the spinning product, is the mean pore area of ​​each pore image block, P is the total number of pore image blocks, kx pmax is the maximum pore area in the pore image block p, kx pmin is the minimum pore area in the pore image patch p.

[0025] As a further solution of the present invention, the convex area statistical unit is used to count the pore image blocks in the concave area, obtain the number of pore image blocks contained in the concave area, the area of ​​each pore image block and the average area of ​​the pore image blocks, and import the number of pore image blocks in the concave area, the area of ​​each pore image block and the average area of ​​the pore image blocks into the concave area defect value calculation formula to calculate the concave area defect value in the spinning product, wherein the concave area defect value calculation formula is:

[0026]

[0027] Where: aq x is the defect value of the concave area, S g is the area of ​​the g-th pore image block in the concave region, G is the total number of pore image blocks in the concave region, is the average area of ​​the pore image blocks in the concave area.

[0028] The technical effects and advantages of a spinning quality visual recognition system of the present invention are as follows: the present invention collects spinning product images through a high-resolution camera, divides the spinning product images into a plurality of image area blocks through grayscale threshold segmentation, and removes the non-shadow area to obtain the shadow area in the spinning product image, distinguishes the shadow area as a convex area or a concave area, and evaluates the quality of the spinning product by constructing a spinning product quality evaluation model, which can effectively solve the problem that the convexity and concavity of the local shadow area cannot be determined during spinning product image recognition, and helps to improve the accuracy and reliability of the evaluation of the spinning quality; can improve production efficiency, reduce labor costs, and at the same time can also reduce subjective errors and improve the consistency and accuracy of the evaluation; is convenient for timely discovery and handling of spinning product quality problems, improves the quality control efficiency on the production line, and reduces the generation of defective products. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A structural schematic diagram of a spinning quality visual recognition system provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0030] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solution in the present invention. Obviously, the described technical solution is only a part of the present invention, not all of it. Based on the technical solution in the present invention, all other technical solutions obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0031] Embodiment 1

[0032] Figure 1 FIG. 1 is a schematic diagram showing the structure of a visual identification system for spinning quality provided by Embodiment 1 of the present invention. Figure 1 As shown, a spinning quality visual recognition system in this embodiment includes a processor, a spinning product image acquisition module and a spinning quality visual recognition module that are connected to the processor in communication.

[0033] The spinning product image acquisition module is used to collect spinning product images through a high-resolution camera;

[0034] The spinning quality visual recognition module is used to visually recognize spinning product images and perform quality assessment.

[0035] The spinning quality visual recognition module includes an image processing unit, a shadow area segmentation unit, an area detection unit and a spinning quality assessment unit;

[0036] The image processing unit is connected to the shadow area segmentation unit and is used for removing noise, enhancing contrast and graying the spinning product image;

[0037] The shadow area segmentation unit is connected to the area detection unit, and is used to segment the spinning product image into a plurality of image area blocks through grayscale threshold segmentation, and remove the non-shadow area to obtain the shadow area in the spinning product image;

[0038] The area detection unit is connected to the spinning quality evaluation unit and is used to distinguish whether the shadow area is a convex area or a concave area;

[0039] The spinning quality evaluation unit is used to evaluate the quality of spinning products by constructing a spinning product quality evaluation model.

[0040] The spinning product image is collected by a high-resolution camera, and the spinning product image is segmented into several image area blocks through grayscale threshold segmentation, and the non-shadow area is eliminated to obtain the shadow area in the spinning product image, and the shadow area is distinguished as a convex area or a concave area. The quality of the spinning product is evaluated by constructing a spinning product quality evaluation model, which can effectively solve the problem that the convexity of the local shadow area cannot be determined during spinning product image recognition, and help to improve the accuracy and reliability of the evaluation of the spinning quality; it can improve production efficiency and reduce labor costs, and at the same time it can also reduce subjective errors and improve the consistency and accuracy of the evaluation; it is easy to timely discover and deal with spinning product quality problems, improve the quality control efficiency on the production line, and reduce the generation of defective products.

[0041] The shadow area segmentation unit includes a non-shadow area recognition subunit, a non-shadow area segmentation subunit and a shadow area extraction subunit, the non-shadow area recognition subunit is connected to the non-shadow area segmentation subunit, and the non-shadow area segmentation subunit is connected to the shadow area extraction subunit;

[0042] The non-shadow area recognition unit is used to recognize the non-shadow area in the spinning product image;

[0043] The non-shadow region segmentation subunit is used for segmenting the non-shadow region in the spinning product image;

[0044] The shadow area extraction unit is used to extract the remaining image area in the spinning product image, and the remaining image area is the shadow area.

[0045] The pixels of the spinning product image have L grayscale levels, and the number of pixels with grayscale level L is L. N , the total number of pixels of the spinning product image is obtained as:

[0046]

[0047] Where: Z N is the total number of pixels of the spinning product image, L is the number of grayscale representations of the pixels of the spinning product image, and L i is the number of pixels with gray level i, LN is the number of pixels whose gray level is represented by L.

[0048] The non-shadow area segmentation subunit is used to segment the non-shadow area in the spinning product image, divide the spinning product image into several pore image blocks according to the specified pore size of spinning, eliminate the pore image blocks in the non-shadow area according to the spinning target grayscale value, and leave the pore image blocks in the shadow area.

[0049] By calculating the grayscale range of the pore image blocks and calculating the mean grayscale range of each pore image block as the spinning target grayscale value, the grayscale range of the pore image blocks is arranged in descending order as a range sequence {K h (i h ,j h ),…,K2(i2,j2),K1(i1,j1)}, where K1(i1,j1) is the pore image block with the smallest grayscale range value, K2(i2,j2) is the pore image block with the second smallest grayscale range value, and K h (i h ,j h ) is the pore image block with the largest grayscale range value; the 40% range sequence {K 0.4h (i 0.4h ,j 0.4h ),…,K2(i2,j2),K1(i1,j1)}, where K 0.4h (i 0.4h ,j 0.4h ) is a pore image block whose grayscale range value is 40% of the maximum value; the grayscale range value of the spinning target grayscale value of the remaining range sequence is compared with the spinning target grayscale value, and the pore image block whose grayscale range value is greater than the spinning target grayscale value is taken as a shadow area and marked.

[0050] By segmenting the spinning product image according to the pore size specified for spinning, the non-shadow area can be accurately segmented, avoiding inaccurate segmentation caused by shadows; by calculating the grayscale range of the pore image block and eliminating the pore image blocks that do not meet the conditions according to the range sequence, the spinning target area, that is, the shadow area, can be effectively extracted; by comparing with the spinning target grayscale value, the pore image block with a grayscale range value greater than the spinning target grayscale value is regarded as the shadow area, which reduces the possibility of misjudgment and improves the accuracy of segmentation; the image segmentation and shadow area extraction can be automatically performed, reducing the need for manual intervention and improving the efficiency and consistency of processing.

[0051] The region detection unit includes a concave-convex region distinguishing unit, a convex region outputting unit and a concave region outputting unit, and the concave-convex region distinguishing unit is connected to the convex region outputting unit and the concave region outputting unit respectively;

[0052] The concave-convex region discrimination unit is used to distinguish the concave-convexity of the shadow region by calculating the grayscale mean of the pore image block in the shadow region, comparing the grayscale value of the pore image block with the grayscale mean;

[0053] The convex region output unit is used to output the pore image block in the convex region;

[0054] The concave region output unit is used to output the pore image block in the concave region.

[0055] The concave-convex area discrimination unit is used to calculate the grayscale mean of the pore image block in the shadow area, and compare the grayscale value of the pore image block with the grayscale mean to distinguish the concave-convexity of the shadow area. The grayscale value of the pore image block is compared with the grayscale mean in the following way:

[0056] when When , the pore image block is a convex area;

[0057] when When , the pore image block is a concave area;

[0058] Among them, s Hq is the gray value of the qth pore image block, is the grayscale mean of the pore image block.

[0059] By calculating the grayscale mean of the pore image block in the shadow area and comparing the grayscale value of the pore image block with the grayscale mean, the concave and convex areas can be automatically distinguished, which is convenient for avoiding the subjectivity and inconsistency of manual judgment and improving the accuracy of concave and convex area discrimination; by comparing the grayscale values, it is possible to quickly determine whether the pore image block is a concave area or a convex area, which speeds up the speed of concave and convex area discrimination and improves processing efficiency; comparing the grayscale value of the pore image block with the grayscale mean can better reflect the grayscale change of the pore image block, thereby more accurately judging the concave and convexity; the judgment of concave and convexity is based on the relationship between the grayscale value and the grayscale mean, rather than a simple threshold comparison, so it can more reliably judge the concave and convexity.

[0060] The spinning quality evaluation unit includes a pore size abnormality analysis unit, a convex region statistics unit and a quality evaluation unit; the pore size abnormality analysis unit and the convex region statistics unit are respectively connected to the quality evaluation unit;

[0061] The pore size anomaly analysis unit is used to analyze the pore size anomaly value of the spinning product through the pore size anomaly analysis calculation model;

[0062] The convex region statistical unit is used to count and integrate the pore image blocks in the concave region;

[0063] The quality evaluation unit is used to evaluate the quality of spun yarn by constructing a spun yarn quality evaluation model.

[0064] The pore size anomaly analysis unit is used to extract the pore area in each pore image block in the spinning product image, arrange the pore area of ​​each pore image block from large to small according to the numerical value, and calculate the average pore area of ​​each pore image block, and import the pore area of ​​each pore image block and the average pore area of ​​each pore image block into the pore size anomaly analysis calculation model to calculate the pore size anomaly value of the spinning product, wherein the calculation formula of the anomaly analysis calculation model is:

[0065]

[0066] Where: yf x is the abnormal value of the pore size of the spinning product, is the mean pore area of ​​each pore image block, P is the total number of pore image blocks, kx pmax is the maximum pore area in the pore image block p, kx pmin is the minimum pore area in the pore image patch p.

[0067] It can effectively extract the pore area of ​​each pore image block in the spinning product image and sort them from large to small according to the area, which helps to intuitively understand the size distribution of different pores; by calculating the mean pore area of ​​each pore image block, the average level of pore area can be obtained, which helps to judge the trend and change of the overall pore size; by importing the pore area and the mean pore area into the abnormal analysis calculation model, the pore size abnormal value can be calculated, which can help identify the abnormal pore size, so as to discover and solve the problem in time; the calculation formula can be used to quantitatively evaluate the abnormal pore size, which helps to accurately locate and quantitatively evaluate the pore size abnormality and provide a basis for subsequent processing.

[0068] The convex area statistical unit is used to count the pore image blocks in the concave area, obtain the number of pore image blocks contained in the concave area, the area of ​​each pore image block and the average area of ​​the pore image blocks, and import the number of pore image blocks in the concave area, the area of ​​each pore image block and the average area of ​​the pore image blocks into the concave area defect value calculation formula to calculate the concave area defect value in the spinning product, wherein the concave area defect value calculation formula is:

[0069]

[0070] Where: aq x is the defect value of the concave area, S g is the area of ​​the g-th pore image block in the concave region, G is the total number of pore image blocks in the concave region, is the average area of ​​the pore image blocks in the concave area.

[0071] By counting the number of pore image blocks in the concave area, the area of ​​each pore image block and the mean area of ​​the pore image blocks, and importing them into the calculation formula for the defect value of the concave area, the defect value of the concave area can be calculated, which is helpful for quantitatively evaluating the quality status of the concave area; by counting the number and area information of the pore image blocks in the concave area, the size distribution and quantity distribution of the concave area can be fully understood, which helps to discover potential defect problems; the mean area of ​​the pore image blocks in the concave area is one of the important indicators to measure the degree of defects in the concave area, and can provide a reference for the calculation of the defect value; through the calculation formula for the defect value of the concave area, the degree of defects in the concave area can be quantified, providing quantitative data support for subsequent quality analysis and improvement.

[0072] The quality assessment unit is used to assess the quality of the spinning product by obtaining the abnormal value of the pore size of the spinning product, the total area of ​​the spinning product image of the defect value of the concave area, and the total area of ​​the G pore image blocks in the concave area, and importing the abnormal value of the pore size of the spinning product, the total area of ​​the spinning product image of the defect value of the concave area, and the total area of ​​the G pore image blocks in the concave area into the spinning product quality assessment model, wherein the formula of the spinning product quality assessment model is:

[0073]

[0074] Where: zl f is the quality index of spinning products, aq x is the defect value of the concave area, yf x is the abnormal value of the pore size of the spinning product, S z is the total area of ​​the spinning product image with defect value in the concave area, S g is the total area of ​​the G pore image blocks in the concave region.

[0075] By introducing the abnormal values ​​of pore size of spinning products, defect values ​​of concave areas, total area of ​​spinning product images and total area of ​​pore image blocks in the concave areas into the spinning product quality assessment model, multiple quality indicators of spinning products are comprehensively considered, which helps to understand the influence of each indicator on the quality of spinning products; by quantitatively evaluating the quality of spinning products, quality problems can be discovered and identified in time, and improvement measures can be determined, which helps to guide adjustments and improvements in the production process and improve the quality level of spinning products; by quantitatively evaluating and analyzing the quality of spinning products, problems can be discovered in time and effective measures can be taken, thereby improving production efficiency and product quality, reducing the rate of defective products and enhancing corporate competitiveness.

[0076] The embodiment of the present invention collects spinning product images through a high-resolution camera, divides the spinning product images into several image area blocks through grayscale threshold segmentation, and removes non-shadow areas to obtain shadow areas in the spinning product images, distinguishes the shadow areas as convex areas or concave areas, and evaluates the quality of the spinning products by constructing a spinning product quality evaluation model. This can effectively solve the problem that the convexity and concavity of the local shadow area cannot be determined during spinning product image recognition, and help to improve the accuracy and reliability of the evaluation of the quality of the spinning products; it can improve production efficiency and reduce labor costs, and at the same time can also reduce subjective errors and improve the consistency and accuracy of the evaluation; it is easy to timely discover and deal with spinning product quality problems, improve the quality control efficiency on the production line, and reduce the generation of defective products.

[0077] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A spinning quality visual recognition system, comprising a processor, a spinning product image acquisition module and a spinning quality visual recognition module which are communicatively connected to the processor, wherein the spinning product image acquisition module is used to collect spinning product images through a high-resolution camera; the spinning quality visual recognition module is used to visually recognize the spinning product images and perform quality assessment, and is characterized in that: The spinning quality visual recognition module includes an image processing unit, a shadow area segmentation unit, an area detection unit and a spinning quality evaluation unit. The spinning quality evaluation unit is used to evaluate the quality of the spinning product by constructing a spinning product quality evaluation model, and obtains the pore size abnormality value of the spinning product, the concave area defect value, the total area of ​​the spinning product image and the concave area defect value. The total area of ​​the pore image blocks, and the pore size abnormality value of the spinning product, the concave area defect value, the total area of ​​the spinning product image, and the concave area defect value are calculated. The total area of ​​the pore image blocks is imported into the spinning product quality evaluation model to evaluate the quality of the spinning product. The formula of the spinning product quality evaluation model is: ； Where: It is an indicator of spinning product quality. is the defect value of the concave area, is the abnormal value of the pore size of the spinning product, is the total area of ​​the spinning product image with defect value in the concave area, In the concave area The total area of ​​the pore image patches.

2. A visual identification system for spinning quality according to claim 1, characterized in that: The image processing unit is used for removing noise, enhancing contrast and graying the spinning product image; The shadow area segmentation unit is used to segment the spinning product image into a plurality of image area blocks through grayscale threshold segmentation, and remove the non-shadow area to obtain the shadow area in the spinning product image; The region detection unit is used to distinguish whether the shadow region is a convex region or a concave region; The spinning quality evaluation unit is used to evaluate the quality of spinning products by constructing a spinning product quality evaluation model.

3. A visual identification system for spinning quality according to claim 1, characterized in that: The shadow area segmentation unit includes a non-shadow area recognition subunit, a non-shadow area segmentation subunit and a shadow area extraction subunit, wherein: The non-shadow area recognition unit is used to recognize the non-shadow area in the spinning product image; The non-shadow region segmentation subunit is used for segmenting the non-shadow region in the spinning product image; The shadow area extraction unit is used to extract the remaining image area in the spinning product image, and the remaining image area is the shadow area.

4. A visual identification system for spinning quality according to claim 3, characterized in that: The non-shadow area segmentation subunit is used to segment the non-shadow area in the spinning product image, divide the spinning product image into several pore image blocks according to the specified pore size of spinning, eliminate the pore image blocks in the non-shadow area according to the spinning target grayscale value, and leave the pore image blocks in the shadow area.

5. A visual identification system for spinning quality according to claim 1, characterized in that: The region detection unit includes a concave-convex region distinguishing unit, a convex region outputting unit and a concave region outputting unit; The concave-convex region discrimination unit is used to distinguish the concave-convexity of the shadow region by calculating the grayscale mean of the pore image block in the shadow region and comparing the grayscale value of the pore image block with the grayscale mean. The grayscale value of the pore image block is compared with the grayscale mean in the following manner: when When , the pore image block is a convex area; when When , the pore image block is a concave area; in, For the The gray value of the pore image block, is the grayscale mean of the pore image block.

6. A visual identification system for spinning quality according to claim 1, characterized in that: The spinning quality assessment unit includes a pore size anomaly analysis unit, a convex area statistical unit and a quality assessment unit. The pore size anomaly analysis unit is used to extract the pore area in each pore image block in the spinning product image, arrange the pore area of ​​each pore image block from large to small according to the numerical value, and calculate the average pore area of ​​each pore image block. The pore area of ​​each pore image block and the average pore area of ​​each pore image block are introduced into the pore size anomaly analysis calculation model to calculate the pore size anomaly value of the spinning product, wherein the calculation formula of the anomaly analysis calculation model is: ； Where: is the abnormal value of the pore size of the spinning product, is the mean pore area of ​​each pore image block, is the total number of pore image patches, is the pore image patch The maximum pore area in is the pore image patch The minimum pore area in .

7. A visual identification system for spinning quality according to claim 6, characterized in that: The convex area statistical unit is used to count the pore image blocks in the concave area, obtain the number of pore image blocks contained in the concave area, the area of ​​each pore image block and the average area of ​​the pore image blocks, and import the number of pore image blocks in the concave area, the area of ​​each pore image block and the average area of ​​the pore image blocks into the concave area defect value calculation formula to calculate the concave area defect value in the spinning product, wherein the concave area defect value calculation formula is: ； Where: is the defect value of the concave area, The concave area The area of ​​the pore image patch, is the total number of pore image patches in the concave region, is the average area of ​​the pore image blocks in the concave area.

Images