A machine vision-based finished tire stress-strain measurement apparatus, method, and program product

Abstract

The present application relates to the technical field of tire machine vision detection, and particularly relates to a kind of finished tire stress strain measurement equipment, method and program product based on machine vision.A kind of finished tire stress strain measurement equipment based on machine vision, the equipment includes tire inflation module, machine vision guiding system, ranging sensor module, measurement control module and computer equipment;The measurement control module is connected with tire inflation module, machine vision guiding system and ranging sensor module, and the working of machine vision guiding system and ranging sensor module is controlled, and the data collected by machine vision guiding system and ranging sensor module is transmitted to computer equipment;The equipment cuts out the cut of the tire surface, inflates the tire, collects the cut photo, and measures the strain value of the cut position using machine vision algorithm.The stress strain measurement is carried out by using the above-mentioned technology.The equipment has the characteristics of high precision and high safety.

Description

Technical Field

[0001] This invention relates to the field of tire machine vision inspection technology, specifically to a machine vision-based finished tire stress and strain measurement device, method, and program product. Background Technology

[0002] Machine vision primarily uses computers to simulate human visual functions, but it is not merely a simple extension of the human eye. More importantly, it possesses some of the functions of the human brain: extracting information from images of objective things, processing and understanding it, and finally applying it to actual detection, measurement, and control. A machine vision system refers to capturing images through machine vision products (i.e., image acquisition devices, which are divided into CMOS and CCD types), transmitting the images to a processing unit, and through digital processing, judging size, shape, color, etc., based on pixel distribution and information such as brightness and color. The system then controls the actions of equipment on-site based on the judgment results.

[0003] The most fundamental characteristic of machine vision systems is their ability to improve production flexibility and automation. In hazardous work environments unsuitable for manual labor or where human vision is insufficient, machine vision is often used to replace manual vision. Machine vision inspection methods can significantly improve production efficiency and automation. Furthermore, machine vision facilitates information integration, making it a foundational technology for computer-integrated manufacturing.

[0004] During actual use, when a tire encounters an obstacle such as a stone, it bulges under stress. Repeated stress can cause cracks to form on the tire sidewall. When the tire pressure reaches a certain level, these cracks become extremely dangerous. Understanding the stress and strain at these cracks is crucial for improving tire manufacturing processes. Manual testing is very dangerous; excessive tire pressure can cause the tire to explode, threatening the lives of workers. Summary of the Invention

[0005] For applications such as stress and strain measurement at cracks in finished tires, this invention provides a machine vision-based stress and strain measurement device for finished tires. This device involves cutting a slit on the tire surface, inflating the tire, capturing images of the slit, and using machine vision algorithms to calculate the strain value at the slit location. By utilizing this technology for stress and strain measurement, the device features high accuracy and high safety.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A machine vision-based stress-strain measurement device for finished tires, the device comprising:

[0008] Tire inflation module, used for inflating tires;

[0009] Machine vision guidance system used to determine the two-dimensional dimensions, diameter, and sidewall pattern of a tire;

[0010] Distance sensor module, used to scan and determine the contour of the tire sidewall;

[0011] Machine vision inspection system is used to identify cracks in tires after they have been cut by a tool and to detect the size of the cracks.

[0012] The measurement and control module is used to control the operation of the tire inflation module, the machine vision guidance system and the distance sensor module, and to transmit the data collected by the machine vision guidance system and the distance sensor module to the computer equipment.

[0013] And a computer device, which acquires the two-dimensional dimensions, diameter, and sidewall pattern of the tire, as well as the outline of the tire sidewall, and generates a cut location distribution map by analyzing the sidewall pattern; and calculates the required strain value based on the detection data by acquiring the size of the crack target before and after tire inflation.

[0014] Preferably, the machine vision guidance system uses a method including the following steps to determine the two-dimensional dimensions, diameter, and sidewall pattern of the tire;

[0015] 1) First, convert the image obtained by the machine vision guidance system into a grayscale image P(x, y);

[0016] 2) Perform edge detection on the grayscale image P(x, y) to obtain the edge image P. e (x,y);

[0017] Define the region of interest (AOI), set the initial rotation angle θ, and give the threshold G;

[0018] Establish a coordinate system with the top left corner of the image as the center, the horizontal axis to the right as the x-axis, and the vertical axis downwards as the y-axis, where (x0, y0) = (0, 0) is the origin. Let x... k The initial horizontal value is 1. The slope m is calculated from the rotation angle θ, and then the vertical value y is calculated. k ;

[0019] At point (x) k ,0) to point (0,y) k Search for two pixels on a straight line that are less than or greater than a threshold G. These will become the corners of a triangle. The search process is performed from left to right of the image. Once the first pixel is established, check if the next 10 adjacent pixels are all less than or greater than the threshold G. If the condition is met, the established pixel is defined as the first corner of the triangle, denoted by (x). i ,y k) is obtained; if not, the next pixel is considered in the same way; the process continues until the first triangle corner is obtained; subsequently, the next pixel established by the continuous scan is defined as the second corner of the triangle, noted (x k ,y i );

[0020] 3) the midpoint N of the first corner (x i ,y k ) and the second corner (x k ,y i ) is calculated, a perpendicular is drawn from the first midpoint N to the other side of the image, the drawing process stops as soon as the perpendicular intersects a pixel less than the threshold G (greater than the threshold G), then the point noted (x p ,y p ) becomes the third corner of the triangle;

[0021] Let A be the length of the straight line from (x i ,y k ) to (x k ,y i ) and B be the length of the straight line from (x i ,y k ) to (x p ,y p ), if A is shorter than B, i.e. the ratio of A to B is less than 1, the iteration process is x k increased by 1 pixel, y k is recalculated, and the previous steps are repeated. If the ratio of A to B is still less than 1, the iteration process is continued until the ratio is equal to 1;

[0022] 4) the second midpoint O is calculated, which is located between the first corner (x i ,y k ) and the third corner (x p ,y p ), and the third midpoint P is calculated, which is located between the second corner (x k ,y i ) and the third corner (x p ,y p );

[0023] 5) a straight line is created from the second midpoint O to the opposite second corner (x k ,y i ) and a straight line is created from the third midpoint P to the opposite first corner (x i ,y k ); the intersection of these two straight lines with the first midpoint N to the opposite third corner (x p ,y p) as the candidate circle center, and collect the detected circle centers into a temporary array;

[0024] 6) change the rotation angle value θ of the next iteration, repeat the previous steps, and select the circle center with the most repetitions from the temporary array as the final circle center;

[0025] 7) use the resulting maximum circle and minimum circle as cutting boundaries, remove the background, and retain the annular region to obtain a tire sidewall image with a sidewall pattern;

[0026] 8) calculate the pixel length of the radius by calculating the distance between the circle center and any angle of the relevant triangle, calibrate the mapping relationship between the pixel length of the tread in the image and the true size, and determine the two-dimensional size and diameter of the tire.

[0027] As a preferred, the distance measuring sensor is a laser displacement sensor.

[0028] As a preferred, the sidewall pattern is analyzed to generate a cut position distribution map, and the specific process is as follows:

[0029] 1) combine individual sidewall patterns in the tire sidewall image to form a sidewall pattern region;

[0030] 2) flip the image pixel grayscale value;

[0031] 3) perform Gaussian blur processing on the tire sidewall image, first calculate the Gaussian convolution kernel, and the calculation formula is:

[0032] 4)

[0033] 5) sample the two-dimensional Gaussian function to obtain a discrete convolution kernel, and also multiply a number so that the sum of all values in the convolution kernel is equal to 1;

[0034] 6) traverse all the pixels of the image, loop twice, and then calculate the pixels near this pixel point in the loop, loop twice;

[0035] 7) use adaptive binarization method to process the image;

[0036] 8) establish a polar coordinate system with the center of the Hough circle transformation as the coordinate center, the angle of the radius tangent to the edge of the sidewall pattern region to the polar axis as θ1, the angle of the other radius tangent to the edge of the sidewall pattern region to the polar axis as θ2, the distance of the sidewall pattern region closest to the center as r1, the distance of the sidewall pattern region farthest from the center as r2, and the coordinates of the sidewall pattern region as:

[0037] G(r, θ) = (r i , θ i )

[0038] Wherein, r1≤r i ≤r2, θ1≤θ i ≤θ2;

[0039] 9) The region coordinates are transmitted to the program to avoid the cutout position distribution map of the sidewall pattern.

[0040] As preferred, the machine vision detection system is used to identify the crack target and detect the target size, and the specific process is as follows:

[0041] 1) The pixel coordinate system of the image is converted into the world coordinate system, and the proportional relationship between the collected image and the real size is calibrated;

[0042] 2) The color image is converted into a gray-scale image by a weighted average method, and the calculation formula is:

[0043]

[0044] Wherein, C′ i is the brightness value of the gray-scale image pixel i position, R i , G i , B i respectively represent the red, green, and blue component values of the color image pixel i position, and T represents the image size;

[0045] 3) Determine the number of all pixel points M*N of the image according to the size of the gray-scale image;

[0046] 4) Traverse each gray value L in the image gray level 0-255, assume that L divides the image into two parts of background and target, determine the number of pixel points and the sum of gray values of the pixel points N0, Sum0 of the background, the number of pixel points and the sum of gray values of the pixel points N1, Sum1 of the target, and calculate the proportions ω0, ω1 of the pixel points of the background and the target in the image, and the calculation formula is:

[0047] ω0=N0 / M*N

[0048] ω1=N1 / M*N

[0049] 5) Calculate the gray mean values μ0, μ1 of the background and the target, and the calculation formula is:

[0050] μ0=Sum0 / N0

[0051] μ1=Sum1 / N1

[0052] 6) Calculate the inter-class variance The calculation formula is:

[0053]

[0054] 7) find the maximum gray value L in the traversal, which is the threshold value for separating the background and the target, and convert the gray image into a binary image;

[0055] 8) extract the crack contour by using the contour extraction method;

[0056] 9) use the extracted crack contour as the initial contour for Sanke curve fitting, and then perform accurate contour positioning by GVF Sanke curve fitting to obtain the crack.

[0057] As a further optimization, the improved Sanke curve fitting is used for accurate contour positioning, and the specific process is as follows:

[0058] 1) use the preliminary contour of the crack region obtained by the contour extraction method as the initial contour line of GVF Snake;

[0059] 2) Snake curve fitting is to change the initial contour to a closed curve with minimum energy by minimizing the energy function, and the energy function is:

[0060]

[0061]

[0062] where the first term is the elastic energy, the second term is the bending energy, and the third term is the external energy;

[0063] 3) GVF Snake curve fitting uses gradient vector flow as external force field, and the energy function is:

[0064]

[0065] 4) where represents the gradient of the edge image;

[0066] 5) solve the GVF field by solving the Euler equation, and the equation is as follows:

[0067]

[0068] 6) where is the Laplace operator.

[0069] As a further optimization, the required strain value is calculated according to the detection data, and the calculation formula of strain is:

[0070] ε = W / 2L

[0071] Where: ε is the strain perpendicular to the cut part; W is the crack growth width of the cut part; L is the length of the cut.

[0072] ​Further, the application further discloses a finished tire stress-strain measurement method based on machine vision.

[0073] Step 1: the tire inflation module inflates the tire;

[0074] Step 2: the machine vision guiding system is used to determine the two-dimensional size, diameter and sidewall pattern of the tire;

[0075] Step 3: the tire is fixed;

[0076] Step 4: the ranging sensor scanning is used to determine the profile of the tire sidewall;

[0077] Step 5: the sidewall pattern is analyzed to generate a cut position distribution map;

[0078] Step 6: the cutter is used to cut the tire;

[0079] Step 7: the machine vision detection system is used to identify the crack target and detect the target size;

[0080] Step 8: the tire is inflated according to the instruction, and steps 7-8 are repeated until the set air pressure is reached;

[0081] Step 9: the required strain value is calculated according to the detection data;

[0082] Step 10: the detection is ended, and the tire is deflated.

[0083] Further, the application further discloses a computer device, including a memory, a processor and a computer program stored in the memory, and the processor executes the computer program to realize steps 2, 5, 7 and 9 in the method.

[0084] Further, the application further discloses a computer program product, including a computer program or instructions, which realizes steps 2, 5, 7 and 9 in the method when executed by the processor.

[0085] The application has the following beneficial effects relative to the prior art: first, the application can automatically measure stress-strain based on machine vision, avoiding manual testing and protecting the safety of workers; second, the machine vision detection algorithm provided by the application can effectively reduce the error of manual measurement, and the algorithm is simple and fast to realize. BRIEF DESCRIPTION OF DRAWINGS

[0086] Figure 1 is a schematic diagram of the structural framework of the equipment of the application.

[0087] Figure 2 is a flowchart of a finished tire stress-strain measurement method based on machine vision.

[0088] Figure 3 is a schematic view of a tire profile taken by a machine vision camera of an embodiment of the present application;

[0089] Figure 4 is a schematic view of a fixed tire;

[0090] Figure 5 is a schematic view of a comparison of a cut position of a tire side with and without a tire side pattern;

[0091] Figure 6 is a schematic view of a tire side cut before and after inflation of a tire. DETAILED DESCRIPTION

[0092] The present application will be further described in conjunction with the accompanying drawings. The present embodiment is implemented on the premise of the technical solution of the present application, and a detailed implementation manner is given, but the protection scope of the present application is not limited to the following embodiments.

[0093] As shown in Figure 1 , the machine vision-based finished tire stress and strain measurement device of the present application comprises:

[0094] a tire inflation module for inflating a tire;

[0095] a machine vision guiding system for determining a two-dimensional size, a diameter, and a tire side pattern of the tire;

[0096] a distance measuring sensor module for scanning and determining a profile of a tire side of the tire;

[0097] a machine vision detection system for identifying a cut target after the tire is cut by a cutter and detecting a size of the target;

[0098] a measurement control module for controlling the tire inflation module, the machine vision guiding system, and the distance measuring sensor module to work, and transmitting data collected by the machine vision guiding system and the distance measuring sensor module to a computer device;

[0099] and the computer device acquires the two-dimensional size, the diameter, and the tire side pattern of the tire, and the profile of the tire side of the tire, generates a cut position distribution map by analyzing the tire side pattern, and calculates a required strain value according to detection data by acquiring the size of the cut target before and after inflation of the tire.

[0100] The machine vision guiding system of the present application comprises a CCD camera 201, and the tire inflation module, a tire holding device, the machine vision detection system, and the distance measuring sensor module are existing functional modules, which will not be described in detail here.

[0101] The embodiment of the application relates to a kind of machine vision-based finished tire stress strain measurement method, which can be used to measure tire crack stress strain.

[0102] As Figure 2 Shown, a kind of machine vision-based finished tire stress strain measurement method, as follows:

[0103] Step 101 tire inflation module inflates tire, so that the inside of tire keeps basic air pressure.

[0104] Step 102 determines the two-dimensional size, diameter of tire, and the specific process is as follows:

[0105] As Figure 3 Shown, CCD camera 201 shoots tire overall profile image, first, the image obtained by CCD camera 201 is converted into gray image P (x, y);

[0106] Edge detection is carried out on gray image P (x, y), and edge image P e (x, y) is obtained;

[0107] Definition area of interest AOI, set initial rotation angle θ, given threshold G;

[0108] With the upper left corner of image as the center of circle, horizontal right as x axis, vertical downward as y axis, establish coordinate system, wherein (x0, y0)=(0,0) is the origin, set x k The horizontal initial value represented is 1, calculate slope m from rotation angle θ, then calculate vertical value y k ;

[0109] Search two pixels less than threshold G (greater than threshold G) on the straight line from point (x k ,0) to point (0,y k ), which will become two triangle angles, and the search process is executed from the left side to the right side of image. Once the first pixel is established, whether the next 10 adjacent pixels are all less than threshold G (greater than threshold G). If the condition is met, the established pixel is defined as the first angle of triangle, expressed as (x i ,y k ). If not, the next pixel is considered in the same way. The process continues until the first triangle angle is obtained. Subsequently, the next pixel established by continuous scanning is defined as the second angle of triangle, denoted as (x k ,y i );

[0110] Calculate the first angle (x i ,y k ) and the second angle (x k ,y iDraw a perpendicular line from the first midpoint N to the other side of the image. The drawing process stops once the perpendicular line intersects a pixel smaller than or greater than the threshold G, and is then marked as (x). p ,y p The point () becomes the third angle of the triangle;

[0111] Let A be a subset of (x) i ,y k ) to (x k ,y i The length of the straight line from (x) to B is the length of the straight line from (x) to (x). i ,y k ) to (x p ,y p If the length of the straight line is x, and A is shorter than B (i.e., the ratio of A to B is less than 1), then the iterative process is x. k Add 1 pixel and recalculate y. k Repeat the previous steps. If the ratio of A to B is still less than 1, continue the iteration process until the ratio equals 1.

[0112] Calculate the second midpoint O, which is located at the first angle (x). i ,y k ) and the third angle (x) p ,y p Between the two points (x, y), calculate the third midpoint P, which lies within the second angle (x, y). k ,y i ) and the third angle (x) p ,y p )between.

[0113] Create a path from the second midpoint O to the second angle (x) of the opposite side. k ,y i The straight line from the third midpoint P to the first angle (x) of the opposite side. i ,y k The two lines are intersected by a line. The intersection of these two lines and the first midpoint N is connected to the third angle (x) of the opposite side. p ,y p The detected straight lines are used as candidate circle centers, and the detected circle centers are collected into a temporary array;

[0114] Change the rotation angle value θ for the next iteration, repeat the previous steps, and select the circle center with the most repetitions from the temporary array as the final circle center;

[0115] Using the obtained largest and smallest circles as cutting boundaries, the background is removed, and the annular area is retained to obtain a tire sidewall image with the sidewall pattern;

[0116] The mapping relationship between the pixel length of the tire tread in the image and the real size is calibrated by calculating the pixel length of the radius between the center of the circle and any angle of the related triangle, and the two-dimensional size and diameter of the tire are determined.

[0117] Step 103 fixes the tire, and the specific process is as follows: as shown in the figure, when the tire reaches the position, the tire holding device 301 holds the tire. Figure 4

[0118] Step 104 scans to determine the profile of the tire, and the specific method is as follows: the laser displacement sensor scans the profile of the tire, and the position and posture of the cutting robot are determined by fitting the appearance through the laser displacement sensor coordinates, so as to ensure that the cutter is perpendicular to the part to be cut.

[0119] Step 105 analyzes the tire side pattern to generate a cutting position distribution map, and the specific method is as follows:

[0120] The single tire side pattern in the tire side image in step 102 is combined to form a tire side pattern area;

[0121] The image pixel gray value is flipped;

[0122] The tire side image is subjected to Gaussian blur processing, and the Gaussian convolution kernel is calculated first, and the calculation formula is as follows:

[0123]

[0124] The two-dimensional Gaussian function is sampled to obtain a discrete convolution kernel, and at the same time, a number is multiplied to make the sum of all values in the convolution kernel equal to 1;

[0125] All pixel points of the image are traversed twice, and then the pixel points near the pixel point are calculated in the loop, and the loop is executed twice;

[0126] The adaptive binary method is used to process the image;

[0127] The polar coordinate system is established with the center of the Hough circle transformation as the coordinate center, the angle of the radius tangent to the edge of the tire side pattern area to the polar axis is θ1, the angle of the other radius tangent to the edge of the tire side pattern area to the polar axis is θ2, the distance of the tire side pattern area closest to the center is r1, the distance of the tire side pattern area farthest from the center is r2, and the coordinates of the tire side pattern area are:

[0128] G(r, θ) = (r i , θ i )

[0129] Wherein, r1≤r i ≤r2, θ1≤θ i ≤θ2.

[0130] ​The region coordinates are transmitted to the program to avoid the sidewall pattern to generate a cut position distribution map.

[0131] Step 106 cuts the tire, and the specific process is as follows: Figure 5 The tire is cut according to the cut position generated in step 105, as shown in steps 401 and 403.

[0132] Step 107 identifies the crack target and detects the target size. The specific process is as follows:

[0133] The machine vision camera captures the tire crack image, converts the pixel coordinate system of the image into the world coordinate system, and calibrates the scale relationship between the collected image and the real size.

[0134] The color image is converted into a gray-scale image by a weighted average method, and the calculation formula is as follows:

[0135]

[0136] wherein C' i is the brightness value of the gray-scale image pixel i position, R i , G i , B i respectively represent the red, green, and blue component values of the color image pixel i position, and T represents the image size.

[0137] The number of all pixel points M*N of the image is determined according to the size of the gray-scale image;

[0138] Each gray value L in the image gray scale 0-255 is traversed, assuming that L divides the image into two parts of background and target, the number of pixel points of the background and the sum of the gray values of the pixels N0, Sum0, the number of pixel points of the target and the sum of the gray values of the pixels N1, Sum1 are determined, and the proportions of the pixel points of the background and the target in the image ω0, ω1 are calculated, and the calculation formula is as follows:

[0139] ω0=N0 / M*N

[0140] ω1=N1 / M*N

[0141] The gray mean values μ0, μ1 of the background and the target are calculated, and the calculation formula is as follows:

[0142] μ0=Sum0 / N0

[0143] μ1=Sum1 / N1

[0144] The inter-class variance is calculated, and the calculation formula is as follows:

[0145]

[0146] The gray value L found in the traversal that maximizes the inter-class variance is the threshold value for separating the background and the target, and the gray image is converted into a binary image;

[0147] The crack contour is extracted by using the contour extraction method;

[0148] The extracted crack contour is used as the initial contour for Sanke curve fitting, and then the GVF Sanke curve fitting is performed to accurately position the contour and obtain the crack;

[0149] The preliminary contour of the crack region obtained by using the contour extraction method is used as the initial contour line of the GVF Snake;

[0150] The Snake curve fitting is to change the initial contour into a closed curve with minimum energy, and the energy function is:

[0151]

[0152] The first term is the elastic energy, the second term is the bending energy, and the third term is the external energy.

[0153] The GVF Snake curve fitting uses the gradient vector flow as the external force field, and the energy function is:

[0154]

[0155] Wherein represents the gradient of the edge image.

[0156] The GVF field is solved by solving the Euler equation, and the equation is as follows:

[0157]

[0158] Wherein is the Laplace operator.

[0159] Step 108, the tire is inflated, and the specific process is as follows: as shown in Figure 6 , the shape of the cut before inflation is 501, the tire is inflated, and after each inflation, it is placed for 24 hours, the inflation pressure is 500kPa, 600kPa, 700kPa, 800kPa, 900kPa respectively, and the shape of the crack expansion is 502, step 106 is repeated to detect the crack, and the inflation is stopped when the set maximum pressure is reached.

[0160] Step 109, the strain value is calculated, and the calculation formula is as follows:

[0161] ε=W / 2L

[0162] Wherein: ε is the strain in the vertical direction of the cut part; W is the fracture growth width of the cut part; and L is the cut length.

[0163] Step 110 tire is deflated.

[0164] The foregoing is a description of embodiments of the application, which enable a person skilled in the art to implement or use the application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and generic principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A machine vision based finished tire stress strain measurement apparatus, characterized by, The device comprises: a tire inflation module for tire inflation; a machine vision guiding system for determining the two-dimensional size, diameter and sidewall pattern of the tire; a distance measuring sensor module for scanning the profile of the tire sidewall; a machine vision detection system for identifying the slit target after the cutter cuts the tire, and detecting the size of the target; a measurement control module for controlling the connection of the tire inflation module, the machine vision guiding system and the distance measuring sensor module, and transmitting the data collected by the machine vision guiding system and the distance measuring sensor module to the computer device; and a computer device for obtaining the two-dimensional size, diameter and sidewall pattern of the tire, generating a cut position distribution map by analyzing the sidewall pattern, and calculating the required strain value according to the size of the slit target before and after the tire is inflated. The machine vision guiding system determines the two-dimensional size, diameter and sidewall pattern of the tire by using a method comprising the following steps: 1) First, convert the image obtained by the machine vision guiding system into a gray image P(x, y); 2) edge detection is performed on the gray-scale image P(x, y) to obtain an edge image P e (x, y); define the area of interest AOI, set the initial rotation angle θ, and give the threshold value G; A coordinate system is established with the upper left corner of the image as the center, the horizontal right as the x axis, and the vertical down as the y axis, where (x0, y0) = (0, 0) is the origin, and x k The horizontal initial value represented by m is 1, the slope m is calculated from the rotation angle θ, and then the vertical value y k ; On the straight line from the point (x k ,0) to the point (0,y k ), search for two pixels less than the threshold G or greater than the threshold G, which will be two triangle corners, the search process is performed from the left to the right of the image; once the first pixel is established, whether the next 10 adjacent pixels are all less than the threshold G or greater than the threshold G; if the condition is met, the established pixel is defined as the first corner of the triangle, denoted as (x i ,y k ); if not, the next pixel is considered in the same way; the process continues until the first triangle corner is obtained; subsequently, the next pixel established by continuous scanning is defined as the second corner of the triangle, denoted as (x k ,y i ); 3) Calculate the midpoint N of the first corner (x i ,y k ) and the second corner (x k ,y i ), draw a perpendicular line from the first midpoint N to the other side of the image, once the perpendicular line intersects a pixel less than a threshold G or greater than a threshold G, the drawing process stops and the point marked as (x p ,y p ) becomes the third corner of the triangle; Let A be the length of the straight line from (x i ,y k ) to (x k ,y i ), B be the length of the straight line from (x i ,y k ) to (x p ,y p ), if A is shorter than B, i.e. the ratio of A to B is less than 1, then the iteration process is x k increased by 1 pixel, y k is recalculated, and the previous steps are repeated; if the ratio of A to B is still less than 1, the iteration process is continued until the ratio equals 1; 4) calculating a second midpoint O between the first corner (x i ,y k ) and the third corner (x p ,y p ), and calculating a third midpoint P between the second corner (x k ,y i ) and the third corner (x p ,y p ); 5) create a straight line from the second midpoint O to the opposite second corner (x k ,y i ) and a straight line from the third midpoint P to the opposite first corner (x i ,y k ); detect the intersection of the two straight lines and the straight line from the first midpoint N to the opposite third corner (x p ,y p ) as a candidate center of the circle, and collect the detected centers of the circle into a temporary array; 6) Change the rotation angle value θ of the next iteration, repeat the previous steps, and select the center of the circle with the most repetitions from the temporary array as the final center of the circle; 7) Take the resulting maximum circle and minimum circle as the cutting boundary, remove the background, and retain the annular region to obtain the tire sidewall image with the sidewall pattern; 8) Calculate the pixel length of the radius by calculating the distance between the center of the circle and any angle of the related triangle, calibrate the mapping relationship between the pixel length of the tread in the image and the true size, and determine the two-dimensional size and diameter of the tire.

2. A machine vision based stress strain measurement apparatus for finished tyres according to claim 1, characterized in that, The distance measuring sensor is a laser displacement sensor.

3. A machine vision based stress strain measurement apparatus for finished tire according to claim 1, wherein, The analysis of the sidewall pattern to generate the cut position distribution map has the following specific process: 1) Combine individual sidewall patterns in the tire sidewall image to form a sidewall pattern area; 2) Flip the image pixel gray value; 3) Perform Gaussian blur processing on the tire sidewall image, first calculate the Gaussian convolution kernel, and the calculation formula is: 4) Sample the two-dimensional Gaussian function to obtain a discrete convolution kernel, and also multiply a number to make the sum of all values in the convolution kernel equal to 1; 5) Traverse all pixel points of the image, loop twice, and then calculate the pixel points near this pixel point in the loop, loop twice; 6) Use adaptive binarization method to process the image; 7) Establish a polar coordinate system with the center of the Hough circle transformation as the coordinate center, the angle of the radius tangent to the edge of the sidewall pattern area to the polar axis as θ1, the angle of the other radius tangent to the edge of the sidewall pattern area to the polar axis as θ2, the distance of the sidewall pattern area closest to the center as r1, the distance of the sidewall pattern area farthest from the center as r2, and the coordinates of the sidewall pattern area as: G(r, θ) = (r i , θ i ) wherein r1≤r i ≤r2, θ1≤θ i ≤θ2; 8) Transfer the region coordinates to the program to avoid generating the cut position distribution map for the sidewall pattern.

4. The machine vision-based stress-strain measurement device for finished tires according to claim 1, characterized in that the machine vision detection system identifies the slit target and detects the size of the target, and the specific process is as follows: 1) Transform the pixel coordinate system of the image into the world coordinate system, and calibrate the proportional relationship between the collected image and the real size; 2) Convert the color image into a grayscale image by using the weighted average method, and the calculation formula is: wherein C i ′ is the luminance value of the gray image pixel i position, R i , G i , B i respectively represent the red, green, blue component values of the color image pixel i position, and T represents the image size; 3) Determine the number of all pixel points M*N of the image according to the size of the grayscale image; 4) Traverse each grayscale value L in the grayscale level 0-255 of the image, assume that L divides the image into two parts of background and target, determine the number of pixel points and the total sum of the grayscale values of the pixel points N0, Sum0 of the background, the number of pixel points and the total sum of the grayscale values of the pixel points N1, Sum1 of the target, and calculate the proportions ω0, ω1 of the pixel points of the background and the target in the image, and the calculation formula is: ω0=N0 / M*N ω1=N1 / M*N 5) Calculate the average grayscale values μ0, μ1 of the background and the target, and the calculation formula is: μ0=Sum0 / N0 μ1=Sum1 / N1 6) Calculate interclass variance The formula is: 7) Find the grayscale value L that makes the inter-class variance maximum in the traversal, and the maximum grayscale value L is the threshold value for dividing the background and the target, and convert the grayscale image into a binary image; 8) Extract the crack contour by using the contour extraction method; 9) Take the extracted crack contour as the initial contour of the Sanke curve fitting, and then perform accurate contour positioning by GVF Sanke curve fitting to obtain the crack.

5. A machine vision-based stress-strain measurement apparatus for finished tyres according to claim 4, characterized in that, The Sanke curve fitting performs accurate contour positioning, and the specific process is as follows: 1) Take the preliminary contour of the crack region obtained by using the contour extraction method as the initial contour line of the GVF Snake; 2) The Snake curve fitting is to change the initial contour into a closed curve with minimum energy by taking the minimization of the energy function as the target, and the energy function is: wherein the first term is the elastic energy, the second term is the bending energy, and the third term is the external energy; 3) The GVF Snake curve fitting takes the gradient vector flow as the external force field to obtain the energy function: 4) wherein denotes the gradient of the edge image; 5) Solve the GVF field by solving the Euler equation, and the equation is as follows: 6) wherein is the Laplacian operator.

6. A machine vision based stress strain measurement apparatus for finished tire according to claim 1, wherein, According to the detection data, the required strain value is calculated, and the calculation formula is: ε=W / 2L Wherein: ε is the strain in the vertical direction of the cut part; W is the crack growth width of the cut part; L is the cut length.

7. A method for measuring stress and strain of a finished tire based on machine vision, characterized in that, The method uses the device of any one of claims 1-6 to perform measurement, and comprises the following steps: Step 1: the tire inflation module inflates the tire; Step 2: the machine vision guidance system determines the two-dimensional size, diameter, and sidewall pattern of the tire; Step 3: fix the tire; Step 4: use the distance measuring sensor to scan and determine the contour of the tire sidewall; Step 5: analyze the sidewall pattern to generate a cut position distribution map; Step 6: cut the tire by using a cutter; Step 7: use the machine vision detection system to identify the crack target and detect the target size; Step 8: inflate the tire according to the instructions, repeat steps 7-8 until the set air pressure is reached; Step 9: calculate the required strain value according to the detection data; Step 10: the detection is completed, and the tire is deflated.

8. A computer device comprising a memory, a processor, and a computer program stored on the memory, wherein the computer program comprises instructions that, when executed by the processor, cause the processor to perform the method of any one of claims 1-7. The processor executes the computer program to realize steps 2, 5, 7, and 9 in the method of claim 7.

9. A computer program product comprising computer programs or instructions, characterized in that, The computer program or instructions implement steps 2, 5, 7 and 9 of the method of claim 7 when executed by the processor.

Images