Intelligent testing method for bulges in automobile tires

Abstract

The present invention relates to the field of image processing technology, and in particular to an intelligent bulge testing method for automobile tires, comprising: obtaining an X-ray detection image of a cord area to obtain multiple detection areas and segmentation thresholds; obtaining the cords in each detection area according to the pixel points segmented under different segmentation thresholds in each detection area; dividing each detection area into two comparison areas, obtaining the symmetry of the cord distribution in each detection area according to the information entropy of the cord edge chain code in each comparison area, obtaining the cord distribution qualification of each detection area, and performing cord distribution qualification detection. The present invention aims to solve the problem that the obtained cord distribution qualification is inaccurate because the grayscale values ​​of the cord pixels in different areas in the X-ray detection image of the tire cord area are relatively different, resulting in the inability to extract the cords in the detection image through a single threshold.

Description

Technical Field

[0001] The invention relates to the technical field of image processing, and in particular to an intelligent bulge testing method applied to automobile tires. Background Art

[0002] As a safety component of a car, tires play a very important role in driving. If a tire fails, the safety of the car cannot be guaranteed. Since a tire bulges, there is a high possibility of a blowout, and a bulge cannot be repaired. Therefore, a bulge test is performed on the finished car tires. The bulge of the tire is caused by a problem with the tire's internal cord layer, so the cord layer in the finished car tire is tested.

[0003] Currently, when testing the cord layer in a tire, an X-ray image of the tire cord layer is obtained, and the distribution of the cords in the X-ray image is used to determine whether the cord layer is qualified. However, since there are steel wire layers and belt layers above the tire cord layer, the grayscale values ​​of different cord pixels on the X-ray image of the cord layer are different, resulting in inaccurate extraction of the cords in the X-ray image of the cord layer through a single threshold, and the results of sparse distribution detection of the cords in the cord layer are not accurate. Summary of the invention

[0004] The invention provides an intelligent bulge testing method applied to automobile tires to solve the existing problems.

[0005] The intelligent bulge testing method for automobile tires of the present invention adopts the following technical solutions:

[0006] The present invention proposes an intelligent bulge testing method for automobile tires, which comprises the following steps:

[0007] Get the cord area Light detection image;

[0008] According to the cord area The gradient value of each pixel in the optical detection image is converted into The light detection image is divided into a plurality of detection areas;

[0009] According to the minimum grayscale value and the maximum grayscale value of the pixel points in each detection area, multiple segmentation thresholds of each detection area are obtained; according to the grayscale variance of the pixel points segmented and the grayscale variance of the pixel points not segmented in each detection area under different segmentation thresholds, the optimal threshold of each detection area is obtained, and then the cord line in each detection area is obtained;

[0010] Each detection area is divided into two comparison areas. According to the information entropy of the cord edge chain code in each comparison area, the cord trend consistency of each comparison area in each detection area is obtained; according to the difference and mean of the cord trend consistency of different comparison areas in the same detection area, and the information entropy of all cord edge chain codes in a detection area, the cord distribution symmetry of each detection area is obtained;

[0011] According to the difference between the cord distribution symmetry of each detection area and the cord distribution symmetry of other detection areas, the cord distribution qualification of each detection area is obtained, and the cord distribution qualification test is performed on the tire to be tested.

[0012] Further, the cord area The gradient value of each pixel in the optical detection image is converted into The light detection image is divided into multiple detection areas, including the following specific methods:

[0013] Get The lateral gradient value of each pixel in the optical detection image is The lateral gradient value in the optical detection image is greater than the preset gradient threshold Pixels of are recorded as edge pixels;

[0014] The line segment obtained by connecting each edge pixel point with the edge pixels in its 8-neighborhood is recorded as the detected edge;

[0015] According to the detected edge, we get Multiple detection areas within the light detection image.

[0016] Furthermore, the method of obtaining multiple segmentation thresholds for each detection area according to the minimum grayscale value and the maximum grayscale value of the pixels in each detection area includes the following specific methods:

[0017] The first The minimum gray value of the pixels in the detection area is taken as the The segmentation threshold with the smallest gray value of the detection area; The maximum gray value of the pixels in the detection area is taken as the The gray value of the segmentation threshold with the largest detection area; the gray value growth step of the preset segmentation threshold is , get the Multiple segmentation thresholds for each detection region.

[0018] Furthermore, the optimal threshold of each detection area is obtained according to the grayscale variance of the pixels segmented and the grayscale variance of the pixels not segmented under different segmentation thresholds in each detection area, including the specific method of:

[0019]

[0020] In the formula, Indicates The segmentation threshold is The probability of the best threshold for the detection area, Indicates The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is: Indicates The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is: Indicates The detection area is The grayscale variance of the pixels not segmented under the segmentation threshold is Indicates The detection area is The grayscale variance of the pixels not segmented under the segmentation threshold is represents the absolute value function, To prevent hyperparameters with denominators equal to 0;

[0021] According to The segmentation threshold is The probability of the best threshold for the detection area, The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is The detection area is The grayscale variance of the pixels not segmented under the segmentation threshold is obtained The optimal threshold for the detection area.

[0022] Further, the obtained The optimal threshold of the detection area includes the following specific methods:

[0023]

[0024] In the formula, Indicates The segmentation threshold is The performance of the best threshold for each detection region, Indicates The segmentation threshold is The probability of the best threshold for the detection area, Indicates The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is: Indicates The detection area is The grayscale variance of pixels that are not segmented under the segmentation threshold;

[0025] Will be the The segmentation threshold with the highest performance of the best threshold of the detection area is taken as the The optimal threshold for the detection area.

[0026] Furthermore, the specific method of obtaining the cords in each detection area includes:

[0027] In the In the detection area, the gray value is less than or equal to The pixel point with the best threshold in the detection area is taken as the cord pixel point;

[0028] The first The line segments obtained by connecting each cord pixel point in the detection area with the cord pixel points in its 8 neighborhoods constitute the first cords within the detection area.

[0029] Furthermore, each detection area is divided into two comparison areas, and the cord trend consistency of each comparison area in each detection area is obtained according to the information entropy of the cord edge chain code in each comparison area, including the specific method of:

[0030] Use a horizontal straight line to The detection area is divided into two areas of equal size, and the Two comparison areas within the detection area;

[0031] Get the The detection area The edge chain code of each cord in the comparison area is obtained. The detection area The information entropy of the edge chain code of each cord in the comparison area;

[0032] Get the In the detection area The calculation formula for the consistency of the cord trend in the comparison area is as follows:

[0033]

[0034] In the formula, Indicates In the detection area The consistency of the cord trend in the comparison area, Indicates In the detection area The number of cords contained in the comparison area, Indicates In the detection area In the comparison area The information entropy of the edge chain code of the cord is Indicates In the detection area The mean information entropy of the edge chain codes of all cords in the comparison area, represents the weight normalization function, represents the absolute value function.

[0035] Furthermore, the cord distribution symmetry of each detection area is obtained according to the difference and mean of the cord trend consistency of different comparison areas in the same detection area, and the information entropy of all cord edge chain codes in a detection area, including the specific method of:

[0036]

[0037] In the formula, Indicates The cord symmetry of the test area, Indicates The consistency of the cord trend in the first comparison area within the detection area, Indicates The cord trend consistency of the second comparison area in the detection area, Indicates The mean value of the cord trend consistency of the two comparison areas in the detection area, Indicates The mean information entropy of the edge chain codes of all cords in the detection area is represents the absolute value function, is an exponential function with a natural constant as its base.

[0038] Furthermore, the cord distribution qualification of each detection area is obtained according to the difference between the cord distribution symmetry of each detection area and the cord distribution symmetry of other detection areas, including the specific method of:

[0039]

[0040] In the formula, Indicates Cord distribution qualification of each test area, Indicates The cord symmetry of the detection area, Indicates The cord symmetry of the detection area, Indicates the number of detection areas. represents the absolute value function, represents the weight normalization function, represents the sigmoid function, is an exponential function with a natural constant as its base.

[0041] Furthermore, the specific method of performing cord distribution qualification test on the tire to be tested includes:

[0042] Preset pass threshold , The cord distribution qualification of the detection area is greater than the qualification threshold , then The cord distribution in the first test area is qualified; if the The cord distribution qualification of the detection area is less than or equal to the qualification threshold , then If the cord distribution in any one of the test areas is unqualified, the cord distribution inside the tire to be tested is unqualified.

[0043] The beneficial effects of the technical solution of the present invention are as follows: the present invention firstly uses the characteristics that the steel wire layer and the belt layer above the cord layer in the tire mainly affect the lateral gradient value of the pixel point of the X-ray detection image of the cord layer through the structural influence of the tire, and according to the lateral gradient value of each pixel point, The optical detection image is divided into multiple detection areas, so that each detection area is more consistently affected by external factors, and the grayscale values ​​of the cord pixels in each detection area are more consistent, and the grayscale values ​​of the non-cord pixels are also more consistent, so that the cords in each detection area are obtained through a single threshold; according to the characteristics that the qualified cords in the tire are all transverse, and the structure of the tire makes the cord distribution in each qualified detection area symmetrical, each detection area is divided into two parts; according to the consistency of all the cord edge chain codes in each part with all the cord edge chain codes in the other part, the symmetry of the cord distribution in each detection area is obtained, so that the symmetry of the cord distribution in each detection area reflects the qualified degree of the cord distribution in the detection area; according to the qualified degree of the cord distribution in each detection area, the result of the cord distribution qualified degree detection of the tire to be inspected is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 It is a flowchart of the steps of the intelligent testing method for bulge of automobile tires applied by the present invention;

[0046] Figure 2 The tire structure diagram is shown below:

[0047] Figure 3 is the image of the tire x-ray detector;

[0048] Figure 4 The tire cord area to be tested Light detection image;

[0049] Figure 5 The tire cord area to be tested Internal region segmentation map of the light detection image;

[0050] Figure 6 The cord area of ​​the tire with unqualified cord distribution Light detection image. DETAILED DESCRIPTION

[0051] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, features and effects of the intelligent bulge testing method for automobile tires proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0052] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0053] The specific scheme of the intelligent bulge testing method for automobile tires provided by the present invention is described in detail below with reference to the accompanying drawings.

[0054] See also Figure 1 , which shows a flowchart of a method for intelligently testing bulges of automobile tires provided by an embodiment of the present invention, the method comprising the following steps:

[0055] Step S001: Obtain the cord area Light detection image.

[0056] Specifically, Figure 2 The diagram shows the tire structure. X-ray detector, to inspect the tire carcass Light detection, get the cord area Light detection image. Among them, tire The image of the X-ray detector is as follows Figure 3 As shown, the cord area Light detection image Figure 4 shown.

[0057] Step S002: According to the cord area The gradient value of each pixel in the optical detection image is converted into The light detection image is divided into multiple detection areas.

[0058] It should be noted that the belt layers covering the cords at different positions inside the tire may be different. Different cord pixels in the optical image are affected differently by the belt layer, resulting in The grayscale values ​​of the cord pixels at different positions in the optical detection image may be different, so the grayscale values ​​of the cord pixels at different positions in the optical detection image may be different. The pixels in the optical detection image are affected by the belt layer. The light detection image is divided into multiple detection areas. Among them, the tire structure diagram is as follows Figure 2 As shown, the belt layer and steel wire layer covering the cords at different positions in the cord layer may be different.

[0059] It should be further explained that in obtaining the cord area When the optical detection image is used, when the other belt layers or steel wire layers covering the cord are less, the belt layers and steel wire layers covering the cord absorb more light. The less light, the The gray value of each cord pixel in the optical detection image reflects the degree to which the cord pixel is affected by other layers above it. The grayscale values ​​of the pixels of the cords affected by the belt layer in the optical detection image are different, and the gradient value of the pixel is determined by the grayscale distribution around the pixel. The gradient value of each pixel in the optical detection image is The light detection image is divided into multiple detection areas.

[0060] It should be further explained that the tire structure diagram and the cord area Light detection image, cord area The light detection image shows the cords in the horizontal direction, while the belt and other layers on the tire have a certain effect on the cord area. The influence of the cords in the optical detection image is north-south, so that the edge of the cord area affected by the belt layer has a larger lateral gradient value. The lateral gradient value of each pixel on the X-ray image divides the X-ray image into multiple detection areas, where Figure 5 For the cord area The cord parts affected by other factors in the light detection image, that is, the framed parts, all represent the cord parts affected by other factors.

[0061] Specifically, use Operator, get The horizontal gradient value of each pixel in the optical detection image. The lateral gradient value in the light detection image is greater than the gradient threshold The pixel points are recorded as edge pixels. The line segments obtained by connecting each edge pixel point with the edge pixels in its 8 neighborhoods are recorded as detected edges. According to the detected edges, we get Multiple detection areas in the light detection image. If there is an edge line segment that does not enclose an area with other edge line segments, the longitudinal gradient value is greater than the gradient threshold. The pixel points are recorded as edge pixels, and multiple new edges are obtained. The detected edge that does not form the detection area is connected to the new edges connected to its two endpoints to obtain the detection area. The operator obtains the horizontal gradient value of each pixel point according to the prior art, which is not described in detail in this embodiment. , this is used as an example for description, and other values ​​may be set in other implementation modes.

[0062] So far, The light detection image is divided into multiple detection areas.

[0063] Step S003: According to the minimum grayscale value and the maximum grayscale value of the pixels in each detection area, multiple segmentation thresholds of each detection area are obtained; according to the grayscale variance of the pixels segmented and the grayscale variance of the pixels not segmented in each detection area under different segmentation thresholds, the optimal threshold of each detection area is obtained, and then the cords in each detection area are obtained.

[0064] It should be noted that in The pixels in a detection area within the optical detection image are similarly affected by the belt layer and other factors, while the cords in the detection area absorb Light. And because the thickness of qualified cords is relatively similar, the grayscale values ​​of cord pixels in the same detection area are relatively similar under ideal conditions, the grayscale values ​​of non-cord pixels are relatively similar, and there is a significant difference in the grayscale values ​​of cord pixels and non-cord pixels. Therefore, multiple thresholds are used to segment a detection area multiple times; according to the grayscale distribution of the segmented pixels and the grayscale distribution of the unsegmented pixels under each threshold, the optimal threshold for each detection area is obtained.

[0065] It should be further explained that, since there is a relatively obvious grayscale difference between the cord pixels and the non-cord pixels in each detection area, when the segmentation threshold reaches the optimal threshold of the detection area, if the threshold continues to be increased, the grayscale variance of the detected pixels in the detection area will have a large increase and the grayscale variance of the undetected pixels will have a small decrease. Therefore, the optimal threshold of each detection area is obtained according to the change of the grayscale variance of the pixels detected in each detection area at different thresholds and the change of the grayscale variance of the pixels not detected.

[0066] It should be further explained that, in actual situations, the grayscale values ​​of the cord pixels in a detection area are within a small grayscale range, and the number of cord pixels at different grayscale values ​​in the grayscale range is different; the grayscale values ​​of non-cord pixels in a detection area also fluctuate within a smaller range, and the number of cord pixels at different grayscale values ​​in the grayscale range is different. This makes it possible that within the grayscale distribution interval of the cord pixels in the detection area and the grayscale distribution area of ​​the non-cord pixels, the closer to the two sides of the interval, the fewer the number of pixels, so that when the segmentation threshold exceeds the optimal threshold and continues to increase, the variance of the grayscale values ​​of the pixels that may be segmented will not increase too significantly, and the variance of the grayscale values ​​of the pixels that are not segmented will not decrease too much. When the segmentation threshold gradually increases from the segmentation threshold with the smallest grayscale value, the variance of the grayscale values ​​of the segmented pixels may not increase much, and the variance of the grayscale values ​​of the unsegmented pixels may not decrease much. Therefore, according to the change of the grayscale variance of the pixels detected at different thresholds in each detection area and the change of the grayscale variance of the pixels not detected, the possibility of each segmentation threshold being the optimal threshold for each detection area is corrected.

[0067] It should be further explained that, due to the large difference in the grayscale values ​​of the cord pixels and the non-cord pixels in each detection area, when the segmentation threshold reaches the optimal segmentation threshold, the sum of the grayscale variance of the segmented pixels and the grayscale variance of the non-segmented pixels is the smallest. Therefore, according to the grayscale variance of the segmented pixels and the grayscale variance of the non-segmented pixels in each detection area under each segmentation threshold, the possibility of each segmentation threshold being the optimal threshold for each detection area is corrected. The optimal threshold for each detection area is obtained. Since the grayscale value of the cord pixels in each detection area is less than the grayscale value of the non-cord pixels, the cords in each detection area are obtained according to the optimal threshold of each detection area.

[0068] Specifically, the The minimum gray value of the pixels in the detection area is taken as the The segmentation threshold with the smallest gray value of the detection area; The maximum gray value of the pixels in the detection area is taken as the The segmentation threshold with the largest gray value in the detection area. Let the gray value growth step of the segmentation threshold be , get the The grayscale growth step size of the segmentation threshold preset in this embodiment is , this is used as an example for description, and other values ​​may be set in other implementation modes.

[0069] Furthermore, using Multiple segmentation thresholds of the detection area are used to The detection area is segmented to obtain the The grayscale variance of the pixels that are not segmented and the grayscale variance of the segmented pixels in the detection area under each segmentation threshold.

[0070] Furthermore, according to The grayscale variance of the pixels segmented in the detection area under multiple segmentation thresholds and the grayscale variance of the pixels not segmented are obtained. The segmentation threshold is The probability of the optimal threshold for a detection area is calculated as follows:

[0071]

[0072] In the formula, Indicates The segmentation threshold is The probability of the best threshold for the detection area, Indicates The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is: Indicates The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is: Indicates The detection area is The grayscale variance of the pixels not segmented under the segmentation threshold is Indicates The detection area is The grayscale variance of the pixels not segmented under the segmentation threshold is represents the absolute value function, To prevent the hyperparameter from having a denominator of 0, this embodiment sets the preset hyperparameter , can be set to other values ​​in other implementations.

[0073] It should be noted that The larger the value of , the more obvious the difference between the grayscale value of the newly added segmented pixel and the grayscale value of the previously segmented pixel is, which further indicates that the possibility that the newly added pixel is a non-cord pixel is greater, that is, the possibility of segmenting the non-cord pixel by increasing the threshold is greater, that is, the first The segmentation threshold is The greater the probability of the optimal threshold for the detection area; The smaller the value of , the greater the possibility of increasing the threshold to segment non-cord pixels, that is, the greater the possibility of increasing the threshold to segment non-cord pixels.

[0074] Further, get the The segmentation threshold is The specific calculation formula for the expression of the optimal threshold of a detection area is as follows:

[0075]

[0076] In the formula, Indicates The segmentation threshold is The performance of the best threshold for each detection region, Indicates The segmentation threshold is The probability of the best threshold for the detection area, Indicates The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is: Indicates The detection area is The grayscale variance of pixels that are not segmented under the segmentation threshold.

[0077] It should be noted that The smaller the value is, the more it conforms to the characteristics that the grayscale values ​​of the cord pixels are close, the grayscale values ​​of the non-cord pixels are close, and the grayscale values ​​of the cord pixels and the non-cord pixels have obvious differences. The segmentation threshold is The greater the probability of an optimal threshold for each detection area.

[0078] Furthermore, The segmentation threshold with the highest performance of the best threshold of the detection area is taken as the The optimal threshold for the detection area. In the detection area, the gray value is less than or equal to The pixel point with the best threshold in the detection area is taken as the cord pixel point. The line segments obtained by connecting each cord pixel point in the detection area with the cord pixel points in its 8 neighborhoods constitute the first cords within the detection area.

[0079] At this point, the cords within each detection area are obtained.

[0080] Step S004: Divide each detection area into two comparison areas, and obtain the cord trend consistency of each comparison area in each detection area according to the information entropy of the cord edge chain code in each comparison area; obtain the cord distribution symmetry of each detection area according to the difference and mean of the cord trend consistency of different comparison areas in the same detection area, and the information entropy of all cord edge chain codes in a detection area.

[0081] It should be noted that since the tire is symmetrical and qualified The qualified cords in the optical detection image are transverse, so that when the cord distribution in a detection area is qualified, if a detection area is divided into two symmetrical and equal parts by a transverse straight line, the distribution of cord pixel points in the two parts is also relatively similar, that is, the edge chain codes of the cords in the two parts are relatively consistent; and when the sparseness of the cord distribution is inconsistent, the edge chain codes of the cords in the two parts are greatly different. Therefore, each detection area is divided into two parts, and the symmetry of the cord distribution in each detection area is obtained according to the consistency of the cord distribution in the two parts. Among them, the image with sparse and inconsistent cord distribution is as follows Figure 6 shown.

[0082] It should be further explained that, since the qualified cords are all transverse, the information entropy of the chain codes of the cords in a detection area is relatively consistent. Therefore, the symmetry of the cord distribution in each detection area is obtained based on the information entropy of the edge chain codes of all cords in each detection area.

[0083] Specifically, first use a horizontal straight line to The detection area is divided into two areas of equal size, and the If it is not possible to use a horizontal straight line to separate the first If the detection area is divided into two equal parts, a horizontal straight line is used to divide the first The detection area is divided into two parts with the closest area. If a cord is divided into two areas in a detection area, then the cord is divided into the area below the transverse straight line.

[0084] Further, get the The detection area The edge chain code of each cord in the comparison area is obtained. The detection area The information entropy of the edge chain code of each cord in the comparison area. The acquisition of the edge chain code of a line segment and the acquisition of the information entropy of an edge chain code are both existing well-known technologies, which will not be described in detail in this embodiment.

[0085] Further, get the In the detection area The calculation formula for the consistency of the cord trend in the comparison area is as follows:

[0086]

[0087] In the formula, Indicates In the detection area The consistency of the cord trend in the comparison area, Indicates In the detection area The number of cords contained in the comparison area, Indicates In the detection area In the comparison area The information entropy of the edge chain code of the cord is Indicates In the detection area The mean information entropy of the edge chain codes of all cords in the comparison area, represents the absolute value function, Represents the weight normalization function, whose input object is the In the detection area The absolute value of the difference between the information entropy of the edge chain code of each cord in the comparison area and the mean information entropy of the edge chain codes of all cords in the comparison area.

[0088] It should be noted that The larger the value, the In the detection area In the comparison area The more the extension direction of a cord is different from that of other cords in the comparison area, the more To determine the value of The weight of In the detection area The edge chain code of each cord in the comparison area is obtained In the detection area When the cord trends of the first and second comparison areas are consistent, the information entropy of the edge chain code of the normal cord smoothes the information entropy of the edge chain code of the abnormal cord, which increases the influence of the abnormal cord. In the detection area The consistency of the cord trends in the contrasting areas is greatly affected by the abnormal cords.

[0089] Further, get the The specific calculation formula for the cord distribution symmetry of a detection area is as follows:

[0090]

[0091] In the formula, Indicates The cord symmetry of the detection area, Indicates The consistency of the cord trend in the first comparison area within the detection area, Indicates The cord trend consistency of the second comparison area in the detection area, Indicates The mean value of the cord trend consistency of the two comparison areas in the detection area, Indicates The mean information entropy of the edge chain codes of all cords in the detection area is represents the absolute value function, is an exponential function with a natural constant as the base. The model shows an inverse proportional relationship. As the input of the model, the implementer can set the inverse proportional function according to the actual situation.

[0092] It should be noted that The larger the value, the The greater the difference in the extension trend of the cords in the two contrasting areas within the first detection area, the more it indicates that The distribution of the cords in the first detection area is less consistent with the characteristic that each detection area in the tire is symmetrical; When the cord trend consistency of each comparison area in the detection area is determined, different weights are assigned to different cord edge chain codes, so that In the detection area The consistency of the cord trend in the contrast area is greatly affected by the abnormal cords, while is to give the same weight to the information entropy of all cords, so that The value reflects the The inconsistency of cord distribution in the detection area.

[0093] Thus, the symmetry of the cord distribution in each detection area is obtained.

[0094] Step S005: according to the difference between the cord distribution symmetry of each detection area and the cord distribution symmetry of other detection areas, the cord distribution qualification of each detection area is obtained, and the cord distribution qualification test is performed on the tire to be tested.

[0095] It should be noted that since the qualified cords are all transverse, and each cord is parallel to other cords under ideal conditions, if the cord distribution in a test area is qualified, then the distribution of the cords in the test area should be relatively consistent with the distribution in other test areas. Therefore, the cord distribution qualification of each test area is obtained through the symmetry of the cords in different test areas.

[0096] Specifically, get the The specific calculation formula for the cord distribution qualification of each test area is as follows:

[0097]

[0098] In the formula, Indicates Cord distribution qualification of each test area, Indicates The cord symmetry of the detection area, Indicates The cord symmetry of the detection area, represents the absolute value function, represents the weight normalization function, whose input object is the cord symmetry of each detection area, Represents the sigmoid function, which is used for normalization processing in this embodiment. Indicates the number of detection areas. is an exponential function with a natural constant as the base. The model shows an inverse proportional relationship. As the input of the model, the implementer can set the inverse proportional function according to the actual situation.

[0099] It should be noted that The larger the value, the The greater the probability of qualified cord distribution in the detection area, the greater the probability of qualified cord distribution in the detection area. The value reflects the The more reliable the cord distribution qualification in each test area is; The larger the value, the The more qualified the cord distribution is within a test area, the greater the probability is.

[0100] Furthermore, a pre-set qualification threshold , The cord distribution qualification of the detection area is greater than the qualification threshold , then The cord distribution in the first test area is qualified. The cord distribution qualification of the detection area is less than or equal to the qualification threshold , then If the cord distribution in any of the detection areas is unqualified, the cord distribution inside the tire to be tested is unqualified, that is, the cord distribution is sparse and inconsistent. , this is used as an example for description, and other values ​​may be set in other implementation modes.

[0101] At this point, this embodiment is completed.

[0102] To sum up, in the embodiments of the present invention, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent testing method for bulges in automobile tires, characterized in that: The method comprises the following steps: Get the cord area Light detection image; According to the cord area The gradient value of each pixel in the optical detection image is converted into The light detection image is divided into a plurality of detection areas; According to the minimum grayscale value and the maximum grayscale value of the pixel points in each detection area, multiple segmentation thresholds of each detection area are obtained; according to the grayscale variance of the pixel points segmented and the grayscale variance of the pixel points not segmented in each detection area under different segmentation thresholds, the optimal threshold of each detection area is obtained, and then the cord line in each detection area is obtained; Each detection area is divided into two comparison areas. According to the information entropy of the cord edge chain code in each comparison area, the cord trend consistency of each comparison area in each detection area is obtained; according to the difference and mean of the cord trend consistency of different comparison areas in the same detection area, and the information entropy of all cord edge chain codes in a detection area, the cord distribution symmetry of each detection area is obtained; According to the difference between the cord distribution symmetry of each detection area and the cord distribution symmetry of other detection areas, the cord distribution qualification of each detection area is obtained, and the cord distribution qualification test is performed on the tire to be tested.

2. The intelligent bulge testing method for automobile tires according to claim 1 is characterized in that: The cord area The gradient value of each pixel in the optical detection image is converted into The light detection image is divided into multiple detection areas, including the following specific methods: Get The lateral gradient value of each pixel in the optical detection image is The lateral gradient value in the optical detection image is greater than the preset gradient threshold Pixels of are recorded as edge pixels; The line segment obtained by connecting each edge pixel point with the edge pixels in its 8-neighborhood is recorded as the detected edge; According to the detected edge, we get Multiple detection areas within the light detection image.

3. The intelligent bulge testing method for automobile tires according to claim 1 is characterized in that: The specific method of obtaining multiple segmentation thresholds for each detection area according to the minimum grayscale value and the maximum grayscale value of the pixel points in each detection area is as follows: The first The minimum gray value of the pixels in the detection area is taken as the The segmentation threshold with the smallest gray value of the detection area; The maximum gray value of the pixels in the detection area is taken as the The segmentation threshold with the maximum gray value in the detection area; The grayscale growth step size of the preset segmentation threshold is , get the Multiple segmentation thresholds for each detection region.

4. The intelligent bulge testing method for automobile tires according to claim 1 is characterized in that: The specific method of obtaining the optimal threshold value of each detection area according to the grayscale variance of the pixels segmented and the grayscale variance of the pixels not segmented under different segmentation thresholds in each detection area is as follows: In the formula, Indicates The segmentation threshold is The probability of the best threshold for the detection area, Indicates The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is: Indicates The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is: Indicates The detection area is The grayscale variance of the pixels not segmented under the segmentation threshold is Indicates The detection area is The grayscale variance of the pixels not segmented under the segmentation threshold is represents the absolute value function, To prevent hyperparameters with denominators equal to 0; According to The segmentation threshold is The probability of the best threshold for the detection area, The detection area is The grayscale variance of the pixel points segmented under the segmentation threshold is The detection area is The grayscale variance of the pixels not segmented under the segmentation threshold is obtained The optimal threshold for the detection area.

5. The intelligent bulge testing method for automobile tires according to claim 4 is characterized in that: The obtained The optimal threshold of the detection area includes the following specific methods: In the formula, Indicates The segmentation threshold is The performance of the best threshold for each detection region, Indicates The segmentation threshold is The probability of the best threshold for the detection area, Indicates The detection area is The grayscale variance of the pixels segmented under the segmentation threshold is: Indicates The detection area is The grayscale variance of pixels that are not segmented under the segmentation threshold; Will be the The segmentation threshold with the highest performance of the best threshold of the detection area is taken as the The optimal threshold for the detection area.

6. The intelligent bulge testing method for automobile tires according to claim 1 is characterized in that: The specific method of obtaining the cords in each detection area includes: In the In the detection area, the gray value is less than or equal to The pixel point with the best threshold in the detection area is taken as the cord pixel point; The first The line segments obtained by connecting each cord pixel point in the detection area with the cord pixel points in its 8 neighborhoods constitute the first cords within the detection area.

7. The intelligent bulge testing method for automobile tires according to claim 1 is characterized in that: The method of dividing each detection area into two comparison areas and obtaining the consistency of the cord trend of each comparison area in each detection area according to the information entropy of the cord edge chain code in each comparison area includes: Use a horizontal straight line to The detection area is divided into two areas of equal size, and the Two comparison areas within the detection area; Get the The detection area The edge chain code of each cord in the comparison area is obtained. The detection area The information entropy of the edge chain code of each cord in the comparison area; Get the In the detection area The calculation formula for the consistency of the cord trend in the comparison area is as follows: In the formula, Indicates In the detection area The consistency of the cord trend in the comparison area, Indicates In the detection area The number of cords contained in the comparison area, Indicates In the detection area In the comparison area The information entropy of the edge chain code of the cord is Indicates In the detection area The mean information entropy of the edge chain codes of all cords in the comparison area, represents the weight normalization function, represents the absolute value function.

8. The intelligent bulge testing method for automobile tires according to claim 1 is characterized in that: The method of obtaining the cord distribution symmetry of each detection area according to the difference and mean of the cord trend consistency of different comparison areas in the same detection area and the information entropy of all cord edge chain codes in a detection area includes: In the formula, Indicates The cord symmetry of the test area, Indicates The consistency of the cord trend in the first comparison area within the detection area, Indicates The cord trend consistency of the second comparison area in the detection area, Indicates The mean value of the cord trend consistency of the two comparison areas in the detection area, Indicates The mean information entropy of the edge chain codes of all cords in the detection area is represents the absolute value function, is an exponential function with a natural constant as its base.

9. The intelligent bulge testing method for automobile tires according to claim 1, characterized in that: The method of obtaining the cord distribution qualification of each detection area according to the difference between the cord distribution symmetry of each detection area and the cord distribution symmetry of other detection areas includes: In the formula, Indicates Cord distribution qualification of each test area, Indicates The cord symmetry of the test area, Indicates The cord symmetry of the test area, Indicates the number of detection areas. represents the absolute value function, represents the weight normalization function, represents the sigmoid function, is an exponential function with a natural constant as its base.

10. The intelligent bulge testing method for automobile tires according to claim 1, characterized in that: The specific method of testing the tire cord distribution conformity to be tested includes: Preset pass threshold , The cord distribution qualification of the detection area is greater than the qualification threshold , then The cord distribution in the first test area is qualified; if the The cord distribution qualification of the detection area is less than or equal to the qualification threshold , then If the cord distribution in any one of the test areas is unqualified, the cord distribution inside the tire to be tested is unqualified.

Images