Tire surface quality judgment method, electronic device and readable storage medium

Abstract

The present disclosure relates to the field of tire inspection, and specifically to a tire surface quality judgment method, electronic device, and readable storage medium. The tire surface quality judgment method includes: obtaining a test image of the tire surface to be tested, performing partition processing and block processing on the tire surface to be tested, confirming the texture complexity of each block image and comparing it with a standard image, thereby judging the tire surface quality. The tire surface quality judgment method disclosed in the present disclosure has higher accuracy, can avoid misjudgment and missed judgment caused by manual observation, and can locate the block containing defects by comparing the texture complexity of each block with the standard image, so as to facilitate the determination of the location of the defect. Through the method disclosed in the present disclosure, the accuracy of tire surface quality judgment can be improved, efficiency can be improved, and manpower can be saved.

Description

Technical Field

[0001] The present disclosure relates to the field of tire detection, and in particular to a tire surface quality judgment method, an electronic device, and a readable storage medium. Background Art

[0002] The quality of automobile tires is closely related to traffic safety. Timely detection of tire defects before leaving the factory can greatly reduce the risk of traffic accidents. Currently, the tire tread is generally scanned by an X-ray tire machine, and the tire quality is judged by manual observation. The current tire quality detection method is prone to missed judgments and misjudgments, and the tire quality detection has low accuracy and low detection efficiency. Summary of the invention

[0003] To overcome the problems existing in the related art, an exemplary embodiment of the present disclosure provides a tire surface quality judgment method, comprising: obtaining an image to be tested of the surface of the tire to be tested according to the tire to be tested; partitioning the image to be tested to obtain a plurality of area images of the image to be tested, wherein the plurality of area images respectively correspond to texture areas in the width direction of the tire to be tested; dividing the plurality of area images into blocks to obtain a plurality of block images of each area image; determining the texture complexity of each of the block images; determining the category to which each of the block images belongs according to the texture complexity of each of the block images, wherein the category is a plurality of texture areas based on a standard image; determining the quality of the tire surface according to the category to which each of the block images belongs and the texture area to which each of the block images corresponds on the tire to be tested; if the category to which the block image belongs is consistent with the texture area to which the block image corresponds, determining that the quality of the texture area on the surface of the tire to be tested corresponding to the block is qualified; if the category to which the block image belongs is inconsistent with the texture area to which the block image corresponds, determining that the quality of the texture area on the surface of the tire to be tested corresponding to the block is unqualified.

[0004] In some embodiments, the partitioning of the image to be tested to obtain multiple regional images of the image to be tested includes: determining the average grayscale value of each column of pixels of the image to be tested according to the image to be tested; partitioning the image to be tested according to the average grayscale value to obtain five rough-cut regional images, corresponding to the left tire bead, left sidewall, crown, right sidewall and right tire bead of the tire to be tested respectively; and obtaining multiple regional images based on the five rough-cut regional images.

[0005] In some embodiments, the division based on the five rough cut region images obtains a plurality of the region images, including: based on the five rough cut region images, taking the center portion of each of the rough cut region images as the region image, wherein the distance between the two ends of the region image in the width direction and the two ends of the corresponding rough cut region image in the width direction is both p / 2; based on the five region images, taking the stitched region image between every two adjacent region images as the region image, wherein the width of the stitched region image is p.

[0006] In some embodiments, determining the texture complexity of each of the block images includes: determining the grayscale co-occurrence matrix of each of the blocks based on the multiple block images; performing normalization processing based on the grayscale co-occurrence matrix to obtain a normalized matrix of each of the blocks; determining the texture complexity of each of the blocks based on the normalized matrix, the texture complexity including one or more of the following texture features: contrast, entropy, energy, and homogeneity.

[0007] In some embodiments, determining the texture complexity of each of the block images includes one or more of the following steps: calculating local binary patterns based on the multiple block images, and determining a histogram of the local binary patterns as the texture complexity of each of the block images; or, performing a wavelet transform based on the multiple block images, and determining texture features in multiple scales and directions as the texture complexity of each of the block images; or, determining the second-order moment of the grayscale value of the block image as the texture complexity of each of the block images based on the multiple block images; or, determining the entropy of the grayscale histogram of the block image as the texture complexity of each of the block images.

[0008] In some embodiments, the multiple categories of the standard image are determined by the following method: based on a standard tire, a standard image of the surface of the standard tire is obtained; the standard image is partitioned to obtain multiple regional standard images of the standard image, wherein the multiple regional standard images respectively correspond to standard texture areas in the width direction of the standard tire; the multiple regional standard images are divided into blocks to obtain multiple block standard images of each regional standard image; the texture complexity of each block standard image is determined; based on the texture complexity of the block standard images, the texture complexity of each regional standard image is determined; the texture complexity of the multiple regional standard images is clustered to determine multiple categories.

[0009] In some embodiments, the partitioning of the standard image to obtain multiple regional standard images of the standard image includes: determining the average grayscale value of each column of pixels in the standard image according to the standard image; partitioning the standard image according to the average grayscale value to obtain five rough-cut regional standard images, corresponding to the left tire bead, left sidewall, crown, right sidewall and right tire bead of the standard tire respectively; based on the five rough-cut regional standard images, taking the center part of each rough-cut regional standard image as the regional standard image, wherein the distance between the two ends of the regional standard image in the width direction and the two ends of the corresponding rough-cut regional standard image in the width direction is p / 2; based on the five regional standard images, taking the spliced ​​regional standard image between every two adjacent regional standard images as the regional standard image, wherein the width of the spliced ​​regional standard image is p.

[0010] In some embodiments, clustering the texture complexity of the plurality of regional standard images to determine a plurality of categories includes: obtaining a standard feature matrix of the standard image based on the texture complexity of the plurality of regional standard images; clustering the standard feature matrix to cluster the regional standard images into five categories, wherein each of the categories corresponds to the tire bead, the junction between the tire bead and the sidewall, the sidewall, the junction between the sidewall and the crown, and the crown of the standard tire.

[0011] In some embodiments, before partitioning the image to be tested to obtain multiple regional images of the image to be tested, one or more of the following steps are included: performing image preprocessing on the image to be tested; or, excluding the background of the image to be tested; or, performing BEMD decomposition on the image to be tested.

[0012] In a second aspect, the present disclosure further provides an electronic device, wherein the electronic device comprises: a memory, a processor, and a data processing program stored in the memory and executable on the processor, wherein the data processing program, when executed by the processor, implements the tire surface quality judgment method as described in any of the above embodiments.

[0013] In a third aspect, the present disclosure further provides a computer-readable storage medium, wherein the computer-readable storage medium stores the following program: the program is used to execute the tire surface quality judgment method as described in any of the above embodiments.

[0014] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure.

[0015] The tire surface quality judgment method provided by the present disclosure can determine the texture complexity of the tire surface and compare it with the standard image, so as to determine whether the tire surface quality is qualified. By zoning and blocking the tire surface image, the texture complexity of each block image is confirmed and compared with the standard image, which has higher accuracy and can avoid misjudgment and missed judgment caused by manual observation. By comparing the texture complexity of each block with the standard image, the block containing defects can be located, which is convenient for determining the location of the defect; by comparing the texture complexity with the standard image to make a quality judgment, this method can be adjusted according to different standard images to meet a variety of different tire detection requirements, and has high flexibility and adaptability. Through the method disclosed in the present disclosure, the accuracy of tire surface quality judgment can be improved, efficiency can be improved, and manpower can be saved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention may be better understood by describing exemplary embodiments of the present invention in conjunction with the accompanying drawings, in which:

[0017] Figure 1 is a flow chart of a tire surface quality judgment method according to an exemplary embodiment of the present disclosure;

[0018] Figure 2 is a flow chart of a tire surface quality judgment method according to another exemplary embodiment of the present disclosure;

[0019] Figure 3 is a flow chart of a tire surface quality judgment method according to another exemplary embodiment of the present disclosure;

[0020] Figure 4 is a flow chart of a tire surface quality judgment method according to another exemplary embodiment of the present disclosure;

[0021] Figure 5 is a flow chart of a tire surface quality judgment method according to another exemplary embodiment of the present disclosure;

[0022] Figure 6 is a flow chart of a tire surface quality judgment method according to another exemplary embodiment of the present disclosure;

[0023] Figure 7 is a flow chart of a tire surface quality judgment method according to another exemplary embodiment of the present disclosure;

[0024] Figure 8 is a flow chart of a tire surface quality judgment method according to another exemplary embodiment of the present disclosure;

[0025] Fig. 9is a flow chart of a tire surface quality judgment method according to another exemplary embodiment of the present disclosure;

[0026] Fig.10 is an image of a tire surface to be tested shown according to an exemplary embodiment of the present disclosure;

[0027] Fig.11 is a flow chart of a BEMD decomposition method according to an exemplary embodiment of the present disclosure;

[0028] Fig.12 is a schematic block diagram of an electronic device according to an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION

[0029] The specific embodiments of the present invention will be described below. It should be noted that in the specific description of these embodiments, in order to provide a concise description, it is impossible for this specification to provide a detailed description of all the features of the actual embodiments. It should be understood that in the actual implementation of any embodiment, just as in the process of any engineering project or design project, in order to achieve the specific goals of the developer and to meet system-related or business-related restrictions, various specific decisions are often made, and this will also change from one embodiment to another. In addition, it can also be understood that although the efforts made in this development process may be complex and lengthy, for ordinary technicians in the field related to the content disclosed by the present invention, some changes such as design, manufacturing or production based on the technical content disclosed in this disclosure are just conventional technical means, and should not be understood as insufficient content of this disclosure.

[0030] Unless otherwise defined, the technical or scientific terms used in the claims and the specification shall have the usual meaning understood by persons with ordinary skills in the technical field to which the invention belongs. The words "first", "second" and similar words used in the patent application specification and the claims of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "One" or "one" and other similar words do not indicate a quantitative limitation, but indicate the existence of at least one. "Include" or "comprises" and other similar words mean that the elements or objects appearing before "include" or "comprises" include the elements or objects listed after "include" or "comprises" and their equivalent elements, and do not exclude other elements or objects. "Connected" or "connected" and other similar words are not limited to physical or mechanical connections, nor are they limited to direct or indirect connections.

[0031] To solve the above problems, Figure 1As shown, the present disclosure provides a tire surface quality judgment method, which may include: step S110 to step S160, and the above steps are specifically described below:

[0032] Step S110, according to the tire to be tested, obtain the image to be tested of the surface of the tire to be tested; according to the tire to be tested, the image to be tested can be obtained by an image acquisition device, wherein the image to be tested can be an image of the tire tread, a part of the tread image of the tire to be tested can be intercepted as the image to be tested, q consecutive rows of pixels in the tread image of the tire to be tested can be taken as the image to be tested, and the image to be tested includes the pattern on the tire tread. The image to be tested can be obtained by an X-ray tire machine, and the image to be tested can be a grayscale image, so as to facilitate the analysis and processing of the image to be tested.

[0033] Step S120, partition the image to be tested to obtain multiple regional images of the image to be tested; wherein the multiple regional images correspond to the texture areas in the width direction of the tire to be tested respectively. According to the image to be tested, the image to be tested can be divided into multiple regional images in the horizontal direction, and the width of each regional image can be different. Since different parts of the tire have different textures, the grayscale value or color of the tire in different parts is very different. The division of the multiple regional images of the image to be tested can be determined by the change of grayscale value or color, so that each regional image can correspond to a different texture area of ​​the tire in the width direction, such as tire bead, sidewall, crown, etc. Since the tread of the tire has different textures at different positions, and the texture difference between different positions is large, the image to be tested can be divided into multiple regional images according to the different texture areas, and the parts with similar texture complexity in the image to be tested can be divided into the same regional image, so that the texture features and texture complexity in each regional image can be distinguished from each other. In the subsequent process of determining the image texture complexity and judging the tire surface quality, the texture complexity error caused by directly judging the image to be tested as a whole can be avoided. Through the partitioning operation, the image to be tested is divided into multiple regional images, and the quality of each regional image is judged independently, which has better accuracy.

[0034] Step S130, dividing the multiple regional images into blocks to obtain multiple block images of each regional image. For the multiple regional images, block processing can be performed so that each regional image is divided into multiple block images of equal size, and the block size of each block image can be p×q, which is convenient for batch processing of the images to be tested.

[0035] Step S140, determining the texture complexity of each block image. For multiple block images, the texture complexity of each block can be determined, and the texture complexity can be represented by texture features such as structure, frequency, information entropy, etc.

[0036] Step S150, according to the texture complexity of each block image, determine the category to which each block image belongs, wherein the category is a plurality of texture areas based on the standard image. The standard image of the defect-free standard tire can be processed in advance, and the standard image can be divided into a plurality of categories according to the different texture areas of the standard tire, each category can correspond to a texture area, and if the texture complexity of a block of the image to be tested belongs to the current category, it can be determined that the block is located on the part of the tire corresponding to the current category. By comparing the texture complexity of each block image with the category of the standard image one by one, it is possible to determine the category to which the texture complexity of each block image belongs on the standard image, that is, the texture area on the surface of the standard tire, and determine the position corresponding to each block on the standard image, so as to facilitate the positioning of the block. Only one standard tire can be processed and classified to judge tires of the same specification or batch with the same texture features, which can avoid the need for image recognition model training. The processing and classification of standard images are simple and fast, and the method of determining the category of the block image of the tire to be tested based on the standard image is more efficient. When judging the surface quality of multiple different tires to be tested, it is only necessary to compare the block image of the tire to be tested with the classification of the corresponding standard tire. This is simple and efficient, and has better universality in the field of tire surface quality judgment.

[0037] Step S160, according to the category to which each block image belongs and the texture area corresponding to each block image on the tire to be tested, the quality of the tire surface is determined; if the category to which the block image belongs is consistent with the texture area corresponding to the block image on the tire to be tested, the quality of the tire surface to be tested is determined to be qualified; if the category to which the block image belongs is inconsistent with the texture area corresponding to the block image on the tire to be tested, the quality of the tire surface to be tested is determined to be unqualified. Whether the tire surface has defects can be determined according to the category to which the block image belongs and the area corresponding to each block, thereby determining whether the quality of the tire is qualified. Since the category corresponding to each block corresponds to different texture areas on the tire surface, the position of the texture complexity of the current block on the defect-free standard tire can be determined according to the category to which each block image belongs, and the regional image corresponding to the block image can determine the position of the block image on the tire to be tested. Therefore, by comparing the category to which each block image belongs and the regional image corresponding to each block image, it can be determined whether the category to which the block image belongs and the region to which the block image belongs correspond to the same parts on the standard tire and the tire to be tested. If they are the same, it can be determined that the surface quality of the tire to be tested is qualified. If they are not the same, the category to which the texture complexity of the block belongs and its corresponding standard tire position are different from the actual one, then it can be determined that the block image is defective, thereby determining that the surface quality of the tire to be tested is unqualified.

[0038] In some embodiments, in step S110, an image of the tire to be tested can be obtained, and q rows of pixels are collected each time as the image to be tested. Steps S120 to S160 are executed on the image to be tested including q rows of pixels. The surface quality of the tire to be tested can be judged in units of q rows of pixels, and the quality of the tire to be tested can be updated in real time. When the category to which the texture complexity of the block image of the image to be tested belongs is detected, its corresponding standard tire position is compared with its corresponding actual tire position. If they are different, it can be directly determined that the surface quality of the tire to be tested is unqualified, and the surface quality judgment is no longer performed on the undetected part of the image of the tire to be tested. In the process of tire quality detection, for the case where it is only necessary to determine whether the tire has defects, it can quickly confirm whether the quality of the tire surface is qualified, thereby improving the efficiency of tire surface quality judgment and having better real-time performance.

[0039] In other embodiments, an image of the tire to be tested may be obtained in step S110, and q rows of pixels are collected each time as the image to be tested, and the image to be tested is processed according to steps S120 to S160. After the processing of the current image to be tested is completed, a new image to be tested is obtained again for image processing and detection, until the complete tread image of the tire to be tested is traversed, and the quality detection of each q rows of the image to be tested is completed, so that the quality of the complete tread of the tire to be tested can be judged, and the quality of each image to be tested is determined by block processing, texture complexity calculation, and determination of the category to which each block image belongs, and finally the quality of all the images to be tested is determined. The quality inspection results of the image can be used to determine whether the quality of the tire to be tested is qualified, and the complete tread and reverse quality of the tire to be tested can be judged. According to the category to which the texture complexity of the block image belongs, the corresponding standard tire position is compared with its corresponding actual tire position. If they are different, it can be determined that the surface quality of the tire to be tested is unqualified, and the specific area where the surface defects of the tire to be tested in the classified image are located can be determined based on the block image. For a tire to be tested with multiple defects, the area where each defect on the tire to be tested is located can be accurately located, which can improve the accuracy of tire surface quality judgment and has better accuracy for situations where the tire defect position needs to be located.

[0040] Through the tire surface quality judgment method of the embodiment of the present disclosure, the image to be tested of the tire to be tested is obtained, the image to be tested is divided into multiple regional images and then divided into multiple blocks again, and the surface quality judgment of each detail area of ​​the tire to be tested can be performed, with higher judgment accuracy. By confirming the texture complexity of each block image, and comparing the texture complexity of each block image with multiple categories of the standard image one by one, it is determined whether the partition to which each block image belongs in the image to be tested is the same as the category to which it belongs in the standard image, thereby confirming whether the texture complexity of each block image meets the texture complexity interval that the standard tire should have in its category, and judging whether the tire surface quality is qualified, the surface quality of the image to be tested can be quantified by texture complexity, which is convenient for judging the surface quality. Since there are many prominent and obvious textures on the tire tread, the texture complexity is used as the basis for judging the tire surface quality, and the tire surface quality can be judged more accurately and precisely. According to the different texture areas of the tire tread, multiple regional images are divided, and for different regional images, the classification of the corresponding standard image is compared according to the texture complexity of each regional image, which can improve the accuracy of tire surface quality judgment. By comparing the texture complexity of each block image with the standard image, higher accuracy can be achieved. The method of this embodiment can be adjusted according to different standard images to adapt to a variety of different tire detection requirements, with high flexibility and adaptability; it can avoid misjudgment and missed judgment caused by manual observation, and can locate the blocks containing defects, which is convenient for determining the defect position on the tire surface, thereby improving the accuracy of tire surface quality judgment, improving efficiency, and saving manpower.

[0041] In some embodiments, Figure 2 As shown, step S120, partitioning is performed according to the image to be tested to obtain multiple area images of the image to be tested, including: step S121, step S122, and step S123.

[0042] Step S121, according to the image to be tested, determine the average gray value of each column of pixels in the image to be tested. According to the gray value of the image to be tested, the gray value of each column of pixels in the image to be tested can be averaged to obtain the average gray value of each column of pixels. The average gray value can be calculated by the following formula:

[0043]

[0044] Where H is the height of the image to be tested, f(x,y) is the grayscale value of the image to be tested at (x,y), and y(x) is an array consisting of the average grayscale values ​​of each column of pixels. By traversing each column of pixels in the image to be tested, the average grayscale value of each column of pixels in the image to be tested can be obtained.

[0045] Step S122, partition the image to be tested according to the average gray value to obtain 5 rough cut area images, such as Fig.10 As shown, they correspond to the left bead A, left sidewall B, crown C, right sidewall D and right bead E of the tire to be tested. The image to be tested can be partitioned according to the average grayscale value and the structure of the tire. Since the tire includes five parts, namely, the left bead A, the left sidewall B, the crown C, the right sidewall D and the right bead E, the grayscale and texture at the positions corresponding to these five parts of the tire are different. The left bead A and the right bead E are darker in color and have larger grayscale values. The left sidewall B and the right sidewall D have relatively small grayscale values, and have fewer textures and are relatively smooth. The crown C has more textures, large grayscale changes, and more obvious texture features, and high texture complexity. The texture difference between the two adjacent regions is large, so that the average grayscale value of the two adjacent regions varies greatly. The image to be tested can be partitioned according to the change and gradient of the average grayscale value of each column of pixels. The image to be tested is divided into five rough-cut area images with the position where the pixel changes greatly as the boundary, so that each rough-cut area image can correspond to five different parts of the surface of the tire to be tested. Since the widths of the left bead A, left sidewall B, crown C, right sidewall D and right bead E are different on the tire, the widths of the five corresponding rough cut areas can also be different. The positions of the bead, sidewall and crown can be determined according to the gradient change of the average gray value, such as Fig.10 As shown, according to the image to be tested, the image to be tested can be divided into left tire bead A, left tire side B, tire crown C, right tire side D and right tire bead E. The width of these five rough cut area images can be expressed as: [a i ,a i+1 )*[c,d],(i=1,2,3,4,5). i ,a i+1 ) represents the position of the i-th rough cut area, a i represents the left end of the i-th region in the width direction, a i+1Indicates the right end of the i-th region in the width direction, and [c, d] indicates the height of the image to be tested. In some cases, the tire to be tested may contain special textures, and the width range of the rough-cut area image can be adjusted according to the distribution of the texture; for some special tires, the crown C may contain a variety of complex textures, such as dense textures in the middle and sparse textures on both sides. In this case, the rough-cut area image corresponding to the crown C can be partitioned according to the distribution of the texture to obtain more than 5 rough-cut area images. Since the image to be tested may be distorted to a certain extent during the image acquisition process, if less than 5 rough-cut area images are obtained through partitioning, the boundaries between the rough-cut areas will be distorted with the image to be tested, resulting in inaccurate partitioning. Through the method of this embodiment, the image to be tested is divided into five rough-cut areas, which can make the features of the five rough-cut areas more obvious, so that each rough-cut area image remains relatively stable in the distortion of the image to be tested, and has higher accuracy.

[0046] Step S123, multiple regional images are obtained based on the 5 rough cut regional images. Since the tire to be tested may be slightly deformed during the image acquisition process, which affects the division of the region and the subsequent quality inspection of the tire by region, each regional rough cut image can be divided again to obtain multiple regional images, thereby avoiding the detection error caused by the rough cut regional image being affected by the tire deformation, which can effectively improve the accuracy of tire surface quality judgment.

[0047] Through step S121, step S122 and step S123, the average grayscale value of each column of the image to be tested is calculated, and the image to be tested is partitioned according to the average grayscale value and the tire surface structure. The image to be tested can be partitioned according to the actual structure of the tire to be tested, so that the image to be tested is divided into a plurality of rough-cut area images corresponding to the actual structures of the tire to be tested, so as to facilitate the subsequent quality judgment according to the area division, avoid the rough-cut area images being affected by tire deformation, and improve the accuracy and efficiency of tire surface quality judgment.

[0048] In some embodiments, Figure 3 As shown, step S123, dividing the five rough-cut region images to obtain a plurality of region images, includes: step S1231 and step S1232.

[0049] Step S1231, based on the five rough cut regional images, the central part of each rough cut regional image is taken as the regional image, wherein the distance between the two ends of the regional image in the width direction and the two ends of the corresponding rough cut regional image in the width direction is p / 2. Since the tire to be tested may be slightly deformed during the image acquisition process, which affects the division of the region and the subsequent quality inspection of the tire by region, the middle part of each regional rough cut image can be selected as the regional image, so as to avoid inaccurate image quality judgment caused by tire deformation, and the distance between the two ends of the regional image in the width direction and the two ends of the corresponding rough cut regional image in the width direction can be p / 2, and the selection of the regional image can be determined according to the following formula:

[0050]

[0051] in, Indicates the width of the i-th region image, a i represents the left end of the i-th region image in the width direction, a i+1 represents the right end of the i-th region image in the width direction, and [c, d] represents the height of the image to be measured.

[0052] Step S1232: based on the five region images, a stitched region image between every two adjacent region images is taken as a region image, wherein the width of the stitched region image is p. The stitched region image between two adjacent region images can be expressed by the following formula:

[0053]

[0054] Among them, a i+1is the boundary between the ith rough cut region and the i+1th rough cut region. For the 5 region images, 4 stitched region images can be determined, and a stitched region image between every two adjacent region images is obtained. The width of the stitched region image is p. The stitched region image can include the ith rough cut region image with a width of p / 2 and the i+1th rough cut region image with a width of p / 2. There are 4 stitched regions in total. The 4 stitched regions can be connected end to end with the 5 region images divided in step S1231. The 4 stitched region images can be determined as region images. A total of 9 regional images can be determined through steps S1231 and S1232, including 5 regional images obtained by dividing the rough cut regional images, which can correspond one-to-one to the five parts of the left bead A, left sidewall B, crown C, right sidewall D and right bead E on the surface of the tire to be tested; the 9 regional images also include 4 regional images determined by the spliced ​​regional images, and the 4 regional images determined by the spliced ​​regional images correspond one-to-one to the junction of the left bead A and the left sidewall B, the junction of the left sidewall B and the crown C, the junction of the crown C and the right sidewall D, and the junction of the right sidewall D and the right bead E, respectively. If more than 5 rough cut regional images are obtained by partitioning in step S122, in step S1231 and step S1232, the central part of each rough cut regional image is taken as the regional image, which may easily lead to overlapping of the selection of regional images. For example, when the first regional image selected is the central part of the first rough cut regional image, and its range is [50,70), the second regional image is the splicing area between the first rough cut area and the second rough cut area. If more than 5 rough cut regional images are selected, the range of the splicing area may be [65,80), so that the splicing area image as the second regional image overlaps with the image of the first regional image, thereby affecting the accuracy of classification. Therefore, based on 5 rough cut regional images, 9 regional images are obtained by division, which has better classification accuracy, thereby improving the accuracy of tire quality judgment. Through step S1231 and step S1232 of this embodiment, 9 regional images can be obtained through division based on 5 rough cut regional images, the rough cut area can be cropped, and the stitching area between the rough cut areas can be recorded as the regional position, so that the number of regional images can be increased. Since the texture features of each stitching area are somewhat different from the texture information in the main part of the rough cut area, using the stitching area image as the regional image can make the division of the regional image more detailed and comprehensive. Through this embodiment, in the case where the image to be tested has defects in the stitching area, the defect position can be located more accurately, and the regional image can be corresponded one by one with each texture area on the surface of the tire to be tested, so that the texture complexity error caused by tire deformation during image acquisition can be eliminated, and the accuracy of the tire surface quality judgment method can be improved.

[0055] In some embodiments, Figure 4As shown, step S140, determining the texture complexity of each block image, may include: step S141, step S142, and step S143.

[0056] Step S141, based on the multiple block images, determine the gray level co-occurrence matrix of each block. The gray level co-occurrence matrix can describe the texture features in the image through the spatial characteristics of the image grayscale, thereby indicating the texture complexity of the image. Therefore, the gray level co-occurrence matrix P can be generated for each block image. i,j (d,θ), thereby determining the texture complexity of the image to be tested, wherein the size of the block image can be p×q, which is determined according to the size of the tire. The gray level of the block is G, and the gray level G is 0 to 255, then the size of the gray level co-occurrence matrix is ​​G×G, and the gray level co-occurrence matrix p i,j In (d,θ), d is the distance and θ is the direction. The distance d and the direction θ are selected according to the actual image requirements in tire detection. Generally, the distance d is selected as 1, and the direction θ is generally selected as 0°, 45°, 90° or 125°. The gray level co-occurrence matrix P of each block is calculated and determined. i,j (d,θ), the gray level co-occurrence matrix can reflect the comprehensive information about the gray level direction, adjacent interval and gray level change amplitude in the block image.

[0057] Step S142, performing normalization processing according to the gray level co-occurrence matrix to obtain a normalized matrix of each block; according to the gray level co-occurrence matrix P i,j (d,θ), perform normalization, that is, divide all elements of the matrix by the sum of all elements of the matrix to obtain a normalized matrix. The normalized matrix is ​​as follows:

[0058]

[0059] The normalized matrix is ​​a probability matrix, where

[0060] Step S143, according to the normalized matrix, the texture complexity of each block is determined, and the texture complexity includes one or more of the following texture features: contrast, entropy, energy, and homogeneity. According to the normalized matrix, one or more texture features can be calculated, and the texture complexity can be represented by the value of one or more texture features. The texture features may include: contrast, entropy, energy, and homogeneity. The texture complexity of the block image can be represented by one or more of the above texture features, and the above texture features can be calculated by the following formula:

[0061]

[0062]

[0063] Among them, contrast can reflect the change of grayscale difference of texture in block image. The higher the contrast, the more significant the texture edge and the clearer the details in block image. Entropy can reflect the randomness of block image. The larger the entropy, the more complex the texture of block image. Energy can reflect the uniformity of texture in block image. The larger the energy, the more uniform the texture of block image. Homogeneity can reflect the similarity of local grayscale in block image. The higher the homogeneity, the smaller the change of grayscale value and the smoother the image.

[0064] Through steps S141 to S143 of this embodiment, by determining the grayscale co-occurrence matrix for each block image, calculating the normalized matrix, and expressing the texture complexity of the block by calculating one or more texture features, the texture complexity can be output in the form of texture feature values ​​through this method, the texture complexity can be quantified, and the texture complexity can be determined more accurately, making the judgment of the tire surface quality more refined and accurate, thereby improving the accuracy of the tire surface quality judgment.

[0065] In some embodiments, Figure 5 As shown, step S140 determines the texture complexity of each block image, including one or more of the following steps: step S144, or step S145, or step S146, or step S147.

[0066] Step S144, based on the multiple block images, calculate the local binary pattern, and determine the histogram of the local binary pattern as the texture complexity of each block image. The local binary pattern histogram of each block image can be determined by calculating the local binary pattern of the pixels in each block and counting the number of occurrences of each pattern. The horizontal axis of the histogram is the possible value of the local binary pattern, that is, the gray value, and the vertical axis is the frequency of occurrence. The histogram of the local binary pattern can be used as a texture feature, and the texture complexity of the block image can be determined therefrom.

[0067] Step S145, performing wavelet transform on the multiple block images, and determining the texture features on multiple scales and directions as the texture complexity of each block image. The block images can be decomposed on different scales by wavelet transform, and the wavelet coefficients on each scale can reflect the spatial frequency characteristics of the texture of the block images at the corresponding scale, thereby determining the texture complexity of the block images. The texture features of the image can also be decomposed in multiple directions by wavelet transform, which can include texture features such as texture structures in horizontal, vertical and diagonal directions, thereby determining the texture complexity of the block images.

[0068] Step S146, based on the multiple block images, determine the second-order moment of the grayscale value of the block image as the texture complexity of each block. A grayscale co-occurrence matrix can be calculated for each block image, and a second-order moment statistical feature can be calculated based on the grayscale co-occurrence matrix, wherein the second-order moment statistical feature may include: energy, contrast, correlation, inverse moment, etc. The above multiple second-order moment statistical features are the texture features of the block image, and the texture complexity of the block image can be determined based on the second-order moment statistical feature.

[0069] Step S147, based on the plurality of block images, the entropy of the grayscale histogram of the block images is determined as the texture complexity of each block. The grayscale histogram of each block image can be calculated based on the block images, wherein the grayscale histogram can reflect the number of pixels of each grayscale level in each block image, and the entropy of the grayscale histogram can be calculated based on the grayscale histogram. The higher the entropy, the more irregular and complex the grayscale distribution is, and vice versa, the lower the entropy, the more uniform or regular the grayscale distribution is, which can be used as a texture feature to determine the texture complexity of the block image.

[0070] The texture complexity can be determined by one or more of the above-mentioned texture features through one or more of steps S144 to S147, so that different texture complexity determination methods can be used for tires to be tested in different situations or different images to be tested, thereby improving the efficiency of tire quality judgment; it is also possible to add one or more of steps S144 to S147 on the basis of grayscale co-occurrence matrix, normalization processing and calculation of texture features to enrich the texture features, thereby improving the accuracy of texture complexity. In the subsequent process of tire quality judgment, the accuracy of tire quality judgment can be effectively improved, and the accuracy of judgment can be improved.

[0071] In some embodiments, Figure 6 As shown, multiple categories of standard images can be determined by the method of steps S210 to S260:

[0072] Step S210, obtaining a standard image of the surface of the standard tire according to the standard tire. The standard image can be obtained by an image acquisition device, or by an X-ray tire machine, wherein the standard image can be an image of the tread of a standard tire, and the standard image includes the pattern on the tread of the tire. The image to be tested can be a grayscale image, so as to facilitate the analysis and processing of the image to be tested. A defect-free standard tire of the same type or batch as the tire to be tested can be selected to obtain the standard image, and the standard tire has the same size, tread pattern and other characteristics as the tire to be tested.

[0073] Step S220, partitioning the standard image to obtain multiple regional standard images of the standard image, wherein the multiple regional standard images correspond to the standard texture areas in the width direction of the standard tire respectively. According to the standard image, the standard image can be divided into multiple regional standard images in the horizontal direction, and the width of each regional standard image can be different. Since different parts of the tire have different textures, the grayscale value or color of different parts of the tire is quite different. The division of multiple regional standard images of the standard image can be determined by the change of grayscale value or color, so that each regional standard image can correspond to different texture areas of the tire in the width direction, such as tire bead, sidewall, crown, etc.

[0074] Step S230, dividing the plurality of regional standard images into blocks to obtain a plurality of block standard images for each regional standard image. The regional standard image can be divided into a plurality of block standard images of the same size by block processing, and the block size of each block standard image can be p×q.

[0075] Step S240, determining the texture complexity of each block standard image; for multiple block standard images, the texture complexity of each block standard image can be determined, and the texture complexity can be represented by texture features such as structure, frequency, and information entropy.

[0076] Step S250, according to the texture complexity of the block standard image, the texture complexity of each regional standard image is determined; according to the texture complexity of multiple block standard images, the texture complexity of multiple block standard images belonging to the same regional standard image can be integrated, and the texture complexity of the entire regional standard image to which they belong is determined by the texture complexity of the multiple block standard images, thereby facilitating the determination of the texture complexity of the texture area of ​​the standard tire surface corresponding to each regional standard image.

[0077] Step S260, clustering the texture complexity of multiple regional standard images to determine multiple categories. A feature vector can be determined for each standard regional image by integrating the texture complexity features of each standard regional image. Each feature vector corresponds to a standard image, and then the multiple regional standard images are divided into multiple categories according to the similarity of the feature vectors through a clustering algorithm. The regional standard images in each category may have similar texture complexity features.

[0078] Through step S210 to step S260, the standard image can be processed, and multiple categories can be determined based on the standard image. Thus, the texture complexity of the qualified tire in different areas is obtained, and a relatively accurate and clear classification is established, which can have a higher judgment accuracy and precision of tire surface quality judgment in the process of judging the tire surface quality of the image to be tested. For a series of identical tires to be tested, before testing the tires to be tested, only one standard tire needs to be processed with a standard image to obtain a standard category, and a series of identical tires can be tested, which has better practicality. Compared with the image detection models such as convolutional neural network models used in the current technology, the standard image detection and clustering processing adopted by the present disclosure, and then the tire to be tested is detected based on the standard image, which can save a lot of computing power, is more convenient and faster, and has higher efficiency.

[0079] In some embodiments, Figure 7 As shown, step S220, partitioning the standard image to obtain multiple regional standard images of the standard image, may include:

[0080] Step S221, according to the standard image, determine the average gray value of each column of pixels in the standard image; according to the gray value of the standard image, the gray value of each column of pixels in the standard image can be averaged to obtain the average gray value of each column of pixels. The average gray value can be calculated by the following formula:

[0081]

[0082] Where H is the height of the standard image, f(x,y) is the grayscale value of the standard image at (x,y), and y(x) is an array consisting of the average grayscale values ​​of each column of pixels. By traversing each column of pixels in the standard image, the average grayscale value of each column of pixels in the standard image can be obtained.

[0083] Step S222, partition the standard image according to the average gray value to obtain 5 rough cut area standard images, such as Fig.10As shown, they correspond to the left bead A, left sidewall B, crown C, right sidewall D and right bead E of the standard tire respectively. The standard image can be partitioned according to the average gray value and the structure of the tire. Since the tire includes five parts, namely, the left bead A, the left sidewall B, the crown C, the right sidewall D and the right bead E, and the texture difference between the two adjacent areas is large, so that the average gray value of the two adjacent areas changes greatly, the standard image can be partitioned according to the change and gradient of the average gray value of each column of pixels. The image to be tested is divided into five rough-cut area standard images with the position where the pixel changes greatly as the boundary, so that each rough-cut area standard image can correspond to five different parts of the standard tire surface one by one. The position of the bead, sidewall and crown can be determined according to the gradient change of the average gray value. According to the image to be tested, the standard image can be divided into the left bead A, the left sidewall B, the crown C, the right sidewall D and the right bead E. The positions of these five rough-cut area standard images can be expressed as: [a i ,a i+1 )*[c,d],(i=1,2,3,4,5). i ,a i+1 ) represents the width of the standard image of the i-th rough cut area, a i represents the left end of the i-th region in the width direction, a i+1 represents the right end of the i-th region in the width direction, and [c, d] represents the height of the standard image.

[0084] Step S223, based on the five rough cut regional standard images, the central part of each rough cut regional standard image is taken as the regional standard image, wherein the distance between the two ends of the regional standard image in the width direction and the two ends of the corresponding rough cut regional standard image in the width direction is p / 2; since the standard tire may be slightly deformed during the image acquisition process, which affects the accuracy of standard image classification, the middle part of each rough cut regional standard image can be selected as the regional standard image, and the distance between the two ends of the regional standard image in the width direction and the two ends of the corresponding rough cut regional standard image in the width direction can be p / 2, and the selection of the regional image can be determined according to the following formula:

[0085]

[0086] in, represents the width of the i-th region image, and [c, d] represents the height of the standard image.

[0087] Step S224, based on the five regional standard images, a stitched regional standard image between every two adjacent regional standard images is taken as the regional standard image, wherein the width of the stitched regional standard image is p. The stitched regional standard image between two adjacent regional standard images can be expressed by the following formula:

[0088]

[0089] Among them, a i+1 is the dividing line between the ith rough cut regional standard image and the i+1th rough cut regional standard image. For the five regional standard images, four stitching regional standard images can be determined, and a stitching regional standard image between every two adjacent regional standard images is obtained. The width of the stitching regional standard image is p. The stitching regional standard image can include the ith rough cut regional standard image with a width of p / 2 and the i+1th rough cut regional standard image with a width of p / 2. There are four stitching standard regions in total. The four stitching standard regions can be connected end to end with the five regional standard images divided in step S1231. The four stitching regional standard images can be determined as regional standard images.

[0090] Through steps S221 to S224, the standard image can be partitioned to obtain 9 regional standard images. By partitioning the standard image, the standard image can be prevented from being affected by the environment and tire deformation during the subsequent clustering of the texture complexity of the standard regional images. This makes the partition of the standard image and its texture complexity less affected by external interference, thereby improving the classification accuracy of the subsequent texture complexity clustering of the standard image, thereby improving the accuracy of the tire surface quality judgment method.

[0091] In some embodiments, Figure 8 As shown, step S240, clustering the texture complexity of multiple regional standard images to determine multiple categories, may include:

[0092] Step S241, according to the texture complexity of multiple regional standard images, a standard feature matrix of the standard image is obtained; the standard feature matrix of the image can be determined according to the texture complexity of the regional standard image confirmed by the texture complexity of each block standard image, and each standard region can include k i blocks, each block can include four different types of texture complexity values, namely: T ij1 ,T ij2 ,T ij3 ,T ij4 , where i represents the i-th region and j represents the j-th block. For the i-th region, a k-th block can be formed. i ×4 data matrix D i, as the texture complexity matrix of the standard image, as shown in the following formula:

[0093]

[0094] And according to the texture complexity matrix, take the mean of each feature, as shown in the following formula:

[0095]

[0096] So we can get the characteristics of region i

[0097] Since there are 9 regional standard images in total, the overall texture complexity features of the 9 regional standard images are integrated into the standard feature matrix D, as shown in the following formula:

[0098]

[0099] Step S242, clustering the standard feature matrix, clustering the regional standard image into 5 categories, wherein each category corresponds to the bead, the junction between the bead and the sidewall, the sidewall, the junction between the sidewall and the crown, and the crown of the standard tire. The standard feature matrix D can be clustered, and the regional standard image can be clustered into 5 categories by using the K-means clustering algorithm, where categories 1 to 5 correspond to the bead, the junction between the bead and the sidewall, the sidewall, the junction between the sidewall and the crown, and the crown of the standard tire, respectively. Fig.10 As shown, the classification of the image from left to right is [1,2,3,4,5,4,3,2,1]. For the image to be tested, the texture complexity of the block image of the image to be tested can be compared with the five categories clustered by the regional standard image to determine the category to which the block image belongs, thereby determining the position of the standard tire to which the block image belongs, and then determining the actual position of the block image on the tire to be tested according to the region of the image to be tested to which the block image belongs. If the actual position of the block image on the tire to be tested is consistent with the position of the standard tire corresponding to the category to which the block image belongs, it can be determined that the surface texture area of ​​the tire to be tested corresponding to the block is qualified. If not, it can be determined that the surface texture area of ​​the tire to be tested corresponding to the block is unqualified, and it can be determined that the block has a defect.

[0100] Through steps S241 and S242, the regional standard images are clustered to determine multiple categories, which can obtain more precise and accurate classification, improve the accuracy of the standard images, and thus improve the accuracy of the tire surface quality judgment method, making it more sensitive to subtle defects on the tire surface, making the tire surface quality judgment more accurate, and less prone to missed judgments and wrong judgments.

[0101] In some embodiments, Fig. 9As shown, in step S120, before partitioning the image to be tested to obtain multiple regional images of the image to be tested, the following may be included: one or more of steps S170 to S190:

[0102] Step S170, image preprocessing is performed on the image to be tested. Image preprocessing may include noise reduction processing, which removes random noise in the image to be tested by methods such as Gaussian filtering, median filtering or bilateral filtering, thereby improving the clarity and consistency of the image. Image preprocessing may also include image correction, which may perform afterglow correction on the image to be tested, thereby improving the clarity of the image. Image preprocessing may also include image enhancement processing, which may enhance the contrast of the image to be tested by methods such as histogram equalization or contrast limited adaptive histogram equalization (CLAHE) on the image to be tested, thereby making the texture information in the image to be tested more obvious.

[0103] Step S180, excluding the background of the image to be tested. The image to be tested may be cut to remove the background of the image to be tested, and only the portion of the image containing the tire tread may be retained. The edge in the image to be tested may be detected by methods such as Canny edge detection or Sobel operator, and then the contour in the image may be found by a contour extraction algorithm, and the largest contour area may be selected as the image to be tested of the tire. Finally, a sub-image of the tire may be cut out according to the bounding box of the contour area, and the background portion may be excluded, thereby reducing the influence of the background on subsequent operations such as extraction of image texture complexity and partitioning.

[0104] Step S190, BEMD decomposition is performed on the image to be tested. The image to be tested can be decomposed by BEMD, and the image features are decomposed into a series of intrinsic mode functions (IMFs), and the first IMF containing the local high-frequency features and main structural information of the image texture is selected for feature extraction, such as Fig.11 As shown in FIG. 1 , the process of performing BEMD decomposition on the image to be tested in step S190 can effectively remove the noise in the image and maintain the texture information of the image through the EMD algorithm.

[0105] Through one or more of steps S170 to S190, the image to be tested can be pre-processed to improve the clarity of the image, reduce noise interference in the image, and make the texture in the image clearer and more prominent, thereby ensuring the clarity and richness of the image texture information, so as to ensure that in subsequent operations, the image texture complexity can be extracted more simply and quickly, thereby improving the efficiency of the tire surface quality judgment method, and making the determined image texture complexity more accurate, thereby improving the accuracy of the tire surface quality judgment method and avoiding the omission of defective tires.

[0106] Based on the same inventive concept, the present disclosure further provides a computer-readable storage medium, which stores the following program, and the program is used to execute the tire surface quality judgment method of any of the aforementioned embodiments.

[0107] like Fig.12 As shown, an embodiment of the present disclosure provides an electronic device 300. The electronic device 300 includes a memory 310, a processor 320, and an input / output (I / O) interface 330. The memory 310 is used to store instructions. The processor 320 is used to call the instructions stored in the memory 310 to execute the tire surface quality judgment method of the embodiment of the present disclosure. The processor 320 is connected to the memory 310 and the I / O interface 330 respectively, for example, through a bus system and / or other forms of connection mechanisms (not shown). The memory 310 can be used to store programs and data, including the program of the tire surface quality judgment method involved in the embodiment of the present disclosure. The processor 320 executes various functional applications and data processing of the electronic device 300 by running the program stored in the memory 310.

[0108] In the embodiment of the present disclosure, the processor 320 can be implemented in at least one hardware form of a digital signal processor (DSP), a field programmable gate array (FPGA), and a programmable logic array (PLA). The processor 320 can be a central processing unit (CPU) or one or a combination of other forms of processing units with data processing capabilities and / or instruction execution capabilities.

[0109] The memory 310 in the embodiment of the present disclosure may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.

[0110] In the embodiment of the present disclosure, the I / O interface 330 may be used to receive input instructions (such as digital or character information, and key signal input related to user settings and function control of the electronic device 300), and may also output various information to the outside (such as images or sounds). In the embodiment of the present disclosure, the I / O interface 330 may include one or more of a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a mouse, a joystick, a trackball, a microphone, a speaker, and a touch panel.

[0111] The present application uses specific words to describe the embodiments of the present application. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or multiple times in different positions in this specification does not necessarily refer to the same embodiment. In addition, some features, structures or characteristics in one or more embodiments of the present application can be appropriately combined.

[0112] In the context of this application, unless the context clearly indicates an exception, the words "a", "an", "a kind" and / or "the" do not refer to the singular, but may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list, and the method or device may also include other steps or elements.

[0113] Similarly, it should be noted that in order to simplify the description of the disclosure of this application and thus help understand one or more application embodiments, in the above description of the embodiments of this application, multiple features are sometimes merged into one embodiment, figure or description thereof. However, this disclosure method does not mean that the features required by the object of this application are more than the features mentioned in the claims. In fact, the features of the embodiments are less than all the features of the single embodiment disclosed above.

[0114] The basic concepts have been described above. Obviously, for those skilled in the art, the above disclosure is only for example and does not constitute a limitation of the present application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements and amendments to the present application. Such modifications, improvements and amendments are suggested in the present application, so such modifications, improvements and amendments still belong to the spirit and scope of the embodiments of the present application.

Claims

1. A method for judging tire surface quality, comprising: According to the tire to be tested, obtaining a test image of the surface of the tire to be tested; Partitioning the image to be tested to obtain a plurality of regional images of the image to be tested, wherein the plurality of regional images respectively correspond to texture regions in the width direction of the tire to be tested; Dividing the plurality of regional images into blocks to obtain a plurality of divided images for each of the regional images; Determining the texture complexity of each of the block images; Determining the category to which each of the block images belongs according to the texture complexity of each of the block images, wherein the category is a plurality of texture regions based on a standard image; Determine the quality of the tire surface according to the category to which each of the block images belongs and the texture area corresponding to each of the block images on the tire to be tested; If the category to which the block image belongs is consistent with the texture area corresponding to the block image, it is determined that the quality of the texture area on the surface of the tire to be tested corresponding to the block is qualified; If the category to which the block image belongs is inconsistent with the texture area corresponding to the block image, it is determined that the quality of the texture area on the surface of the tire to be tested corresponding to the block is unqualified; Wherein, determining the texture complexity of each of the block images includes: Determine a gray level co-occurrence matrix of each of the blocks according to the plurality of block images; Performing normalization processing according to the gray level co-occurrence matrix to obtain a normalized matrix of each of the blocks; Determining the texture complexity of each of the blocks according to the normalized matrix, wherein the texture complexity includes one or more of the following texture features: contrast, entropy, energy, and homogeneity; The multiple categories of the standard image are determined by the following method: According to a standard tire, obtaining a standard image of the surface of the standard tire; Partitioning the standard image to obtain a plurality of regional standard images of the standard image, wherein the plurality of regional standard images respectively correspond to standard texture areas in a width direction of the standard tire; Dividing the plurality of regional standard images into blocks to obtain a plurality of divided standard images for each of the regional standard images; Determining the texture complexity of each of the block standard images; Determining the texture complexity of each of the regional standard images according to the texture complexity of the block standard images; Clustering the texture complexity of a plurality of standard images of the region to determine a plurality of categories; The step of clustering the texture complexity of the plurality of regional standard images to determine a plurality of categories includes: Obtaining a standard feature matrix of the standard image according to the texture complexity of the plurality of standard images of the region; The standard feature matrix is ​​clustered so that the regional standard image is clustered into five categories, wherein each category corresponds to the bead, the junction between the bead and the sidewall, the sidewall, the junction between the sidewall and the crown, and the crown of the standard tire.

2. The tire surface quality judgment method according to claim 1, wherein: Partitioning the image to be tested to obtain a plurality of regional images of the image to be tested includes: According to the image to be tested, determining the average grayscale value of each column of pixels of the image to be tested; According to the average gray value, the image to be tested is partitioned to obtain five rough cut area images, which respectively correspond to the left bead, left sidewall, crown, right sidewall and right bead of the tire to be tested; A plurality of the region images are obtained based on the five rough cut region images.

3. The tire surface quality judgment method according to claim 2, wherein: The step of obtaining a plurality of the regional images based on the five rough cut regional images includes: Based on the five rough cut region images, taking the central part of each rough cut region image as the region image, wherein the distance between the two ends of the region image in the width direction and the two ends of the corresponding rough cut region image in the width direction is p / 2; Based on the five region images, a stitched region image between every two adjacent region images is taken as the region image, wherein the width of the stitched region image is p.

4. The tire surface quality judgment method according to claim 1, wherein: Determining the texture complexity of each of the block images comprises one or more of the following steps: According to the plurality of block images, a local binary pattern is calculated, and a histogram of the local binary pattern is determined as the texture complexity of each of the block images; or, According to the plurality of block images, wavelet transform is performed to determine the texture features in a plurality of scales and directions as the texture complexity of each of the block images; or, According to the plurality of block images, determining the second-order moment of the grayscale value of the block image as the texture complexity of each block; or, According to the multiple block images, entropy of the grayscale histogram of the block images is determined as the texture complexity of each block.

5. The tire surface quality judgment method according to claim 1, wherein: Partitioning the standard image to obtain a plurality of regional standard images of the standard image includes: According to the standard image, determining the average grayscale value of each column of pixels in the standard image; According to the average gray value, the standard image is partitioned to obtain 5 rough cut area standard images, which respectively correspond to the left bead, left sidewall, crown, right sidewall and right bead of the standard tire; Based on the five rough cut area standard images, taking the central part of each rough cut area standard image as the area standard image, wherein the distance between the two ends of the area standard image in the width direction and the two ends of the corresponding rough cut area standard image in the width direction is p / 2; Based on the five regional standard images, a stitched regional standard image between every two adjacent regional standard images is taken as the regional standard image, wherein the width of the stitched regional standard image is p.

6. The tire surface quality judgment method according to claim 1, wherein: Before partitioning the image to be tested to obtain a plurality of regional images of the image to be tested, one or more of the following steps are included: Performing image preprocessing on the image to be tested; or, Excluding the background of the image to be measured; or, Perform BEMD decomposition on the image to be tested.

7. An electronic device, wherein: The electronic device comprises: a memory, a processor and a data processing program stored in the memory and executable on the processor, wherein the data processing program, when executed by the processor, implements the tire surface quality judgment method as described in any one of claims 1 to 6.

8. A computer-readable storage medium, wherein the computer-readable storage medium stores the following program: The program is used to execute the tire surface quality judgment method as described in any one of claims 1-6.

Images