Tire bubble X-ray feature recognition device and method based on curved surface geometry

Abstract

The invention discloses a tire bubble X-ray feature recognition device and method based on curved surface geometry, and belongs to the technical field of tire X-ray bubble detection. The identification device comprises the following parts. The X-ray machine photograph data preprocessor is deployed at the edge end and is used for preprocessing the acquired tire X-ray machine photograph; the data storage center is deployed at a cloud end and is used for centrally storing the preprocessed X-ray machine photo data and subsequent calculation result data; the data calculation center is deployed at a cloud end, introduces a curved surface geometric analysis method, and obtains bubble characteristics from a tire X-ray machine photo sample; and the bubble feature display and statistical system is used for visually displaying the bubbles identified from the tire X-ray photograph sample and statistical data related to the bubble features.

Description

Technical Field

[0001] The present application relates to the technical field of tire X-ray bubble detection, and more specifically, to a tire bubble X-ray feature recognition device and method based on curved surface geometry. Background Art

[0002] Tires are a key component of automobiles, and their quality directly affects driving safety. During the tire production process, internal defects such as bubbles may occur due to factors such as materials and processes. If these defects are not discovered in time, they will seriously affect the performance and service life of the tire. Given that X-rays have excellent penetrating properties and can be used to detect the internal structure of tires and evaluate product quality, tire X-ray inspection technology came into being.

[0003] Traditional tire X-ray inspection relies on manual identification, which is inefficient and prone to errors. With the development of image processing and pattern recognition technology, tire X-ray defect detection technology has begun to develop in the direction of automation and intelligence. Among them, deep learning technology has significantly improved the accuracy and speed of defect identification. Domestic and foreign researchers have proposed a variety of algorithms and models, but there is currently no method that can fully meet all the needs of industrial production. Domestic tire manufacturers mainly rely on imported X-ray inspection equipment, but these devices are not only expensive, but may not fully adapt to the specific needs of the domestic production environment.

[0004] The internal structure of tires is complex, and the number of layers and arrangement of steel cords vary greatly, resulting in complex textures in X-ray images, which increases the difficulty of feature extraction and limits the performance of bubble recognition algorithms. There are differences in defect standards and classifications for tire products at home and abroad, resulting in adaptability issues for defect detection systems developed abroad in actual domestic applications. Most domestic tire manufacturers still rely on manual inspection, which carries the risk of false detection and missed detection, and low inspection efficiency.

[0005] In summary, although tire X-ray bubble detection technology has made some progress, it still faces technical challenges and practical application difficulties. Among them, extracting tire bubble features based on X-ray images is a key basic link, and innovative technologies are needed to overcome existing difficulties. Improving the accuracy, adaptability and efficiency of the detection algorithm, as well as developing a detection system suitable for the domestic production environment, will be the focus of future research. Summary of the invention

[0006] In order to overcome a series of defects in the prior art, the purpose of this application is to provide a tire bubble X-ray feature recognition method based on curved surface geometry for the above problems, which specifically includes the following steps:

[0007] S1. On the tire X-ray electronic photograph, mark the bubble area, take the upper left corner vertex as the coordinate origin, and define a three-dimensional vector (x, y, z) at each pixel point, where x represents the horizontal coordinate of the pixel, y represents the vertical coordinate of the pixel, and z represents the gray value of the pixel. The larger the z value, the higher the brightness of the corresponding pixel point on the picture. All (x, y, z) constitute a set P, the three-dimensional space where P is located is recorded as U, and the horizontal and vertical coordinates constitute the projection coordinate system V;

[0008] S2. In the projection coordinate system V, the distance between the centers of two adjacent pixels along the horizontal axis is d x , the distance along the ordinate axis is denoted as d y ;

[0009] S3, select a suitable positive integer C as an adjustable parameter;

[0010] S4, in the projection coordinate system V, with d min-y =C*d y Length is the interval, and several straight lines are drawn along the horizontal axis. min-x =C*d x Length is the interval, and several straight lines are drawn along the ordinate direction to grid the projection coordinate system V. Each grid Project P upward along the gray value to grid it and get several p i-j ,in, The four vertices are:

[0011]

[0012]

[0013]

[0014]

[0015] S5, fitting P to obtain the gray value surface s and its equation in the three-dimensional space;

[0016] S6, any The rectangular area formed is recorded as The inscribed ellipse is denoted by

[0017] S7, and Mapped onto the gray value surface s, we get the mapping structure and With p i-j correspond;

[0018] S8, calculate this time The characteristic parameter group And calculate The characteristic parameter group

[0019] S9, obtain a large number of tire X-ray electronic photo bubble area samples, and for each type of bubble, calculate its and Get the bubble type and distributed;

[0020] S10, the bubble feature extraction calculation based on the tire X-ray photo is completed.

[0021] Through the above technical scheme, firstly, a sample of tire bubble X-ray machine photo is obtained, and based on the three-dimensional space composed of the photo plane and gray value, a quadratic polynomial is used to approximate the gray value distribution surface. Then, a rectangle and an ellipse inscribed in the rectangle are constructed on the photo plane, and this pair of figures is reversely mapped to the gray value surface. The rectangle and the ellipse will deform with the geometric characteristics of the gray value surface to form a pair of new surface figures. The characteristics of the tire bubble area are quantified by measuring and comparing the deformation of geometric characteristics such as angle, perimeter, area, and curvature. The advantages of this method are: fully considering the geometric elements and their relationship, as well as their deformation measurement from plane geometry to surface geometry, grasping the bubble geometric characteristics from the microscopic characteristics, and then quantifying them into bubble characteristic indicators, with good versatility and adaptability, enriching the content of bubble feature analysis, and improving the feature distinction ability. At the same time, the calculation process is simple and the deployment is flexible.

[0022] Optionally, in step S5, the steps of calculating the surface s equation are as follows:

[0023] S5-1. Set a binary quadratic function to be determined: f(x, y) = ax 2 +bxy+cy 2 +d;

[0024] S5-2, the solution P has more than 4 points, that is, (x k ,y k , z k ), k=1, 2, 3,...N, N≥4;

[0025] S5-3, construct the following formula:

[0026] S5-4. Order:

[0027] S5-5. Solve the above four equations together to obtain a, b, c, d, which is the equation of the surface s: z = f (x, y) = ax 2 +bxy+cy2 +d.

[0028] Through the above technical solution, the gray value surface of the tire bubble X-ray machine photo is obtained, which forms the basis for the subsequent mapping of the plane to the surface and the introduction of surface geometry calculation. It is particularly important to point out that the gray value surface is fitted with a binary quadratic function here to demonstrate the basis of the core method of surface geometry. In fact, the gray value surface can also be fitted with a cubic or even higher function, or the Fourier approximation method can be introduced to fit the gray value surface to obtain a fitting effect that meets other scenarios. The advantage of this solution is that the calculation is simple and the results are intuitive and clear.

[0029] Optionally, in step S6, Calculate the inscribed ellipse The steps are as follows:

[0030] S6-1. Acquisition The four vertices of:

[0031]

[0032]

[0033]

[0034]

[0035] S6-2. Acquisition The center point:

[0036] S6-3, Upper inscribed ellipse The equation is:

[0037] Through the above technical solution, the ellipse equation expression based on the tire X-ray machine photo is obtained. Among them, d min-x and d min-y The two variables are based on the pixel distance d x and d y The advantage of the structure is that it can adapt to the measurement scale requirements of different types of X-ray machine photo analysis.

[0038] Optionally, in step S7, calculate and The steps are as follows:

[0039] S7-1. Obtain the equation of the surface s: z = f (x, y) = ax 2 +bxy+cy 2 +d;

[0040] S7-2, Calculation The steps are:

[0041] The equation of the projection of the first edge of on V is: y = j * d min-y , the two endpoints are: Substituting into the equation of surface s, we get The equation of the first side: z = ax 2 +b*j*d min-y *x+c*j 2 *d min-y 2 +d, where x∈[i*d min-x , (i+1)*d min-x ];

[0042] Using a similar principle, we can calculate The equations of the other three sides are:

[0043] z=a*(i+1) 2 *d min-x 2 +b*(i+1)*d min-x*y +cy 2 +d, where

[0044] y∈[j*d min-y , (j+1)*d min-y ];

[0045] z=ax 2 +b*(j+1)*d min-y *x+c*(j+1) 2 *d min-2 +d, where x∈[i*d min-x , (i+1)*d min-x ];

[0046] z=a*i 2 *d min-x 2 +b*i*d min-x *y+cy 2 +d, where y∈[j*d min-y , (j+1)*d min-y ];

[0047] S7-3. Calculation The steps are as follows:

[0048] Obtaining the inscribed ellipse based on S6-3 Parametric equation: x = 0.5*cosθ*d min-x +(i+0.5)*d min-x, y=0.5*sinθ*d min-y +(j+0.5)*d min-y ;

[0049] Substitute the above parametric equation into z = f (x, y) = ax 2 +bxy+y 2 +d, we get the surface s The parametric equations are:

[0050] z=a(0.5*cosθ*d min-x +(i+0.5)*d min-x ) 2 +b*(0.5*cosθ*d min-x +(i+0.5)*d min-x )*(0.5*sinθ*d min-y +(j+0.5)*d min-y )+c(0.5*sinθ*d min-y +(j+0.5)*d min-y ) 2 +d.

[0051] Through the above technical scheme, the rectangle and its inscribed ellipse constructed on the photo plane are mapped to the gray value surface, and the equation expression of the mapped figure or the equations of its components are obtained. The calculation adopts the conventional transformation calculation method and integrates the pixel requirements of the tire X-ray machine photo, so the calculation result is reliable and has good adaptability.

[0052] Optionally, in step S8, calculate and The steps are as follows:

[0053] S8-1. Cause It is a collection of characteristic parameters. The representative parameters need to be obtained according to the following steps. The calculation method is as follows;

[0054] S8-1-1. Calculation Circumference and Perimeter ratio Here are the steps:

[0055] perimeter The calculation formula is:

[0056]

[0057] therefore,

[0058] S8-1-2. Calculation Area and Area ratio Here are the steps:

[0059] area The calculation formula is:

[0060]

[0061] therefore,

[0062] S8-1-3. Calculation sum of interior angles and Ratio of the sum of the interior angles Here are the steps:

[0063] sum of interior angles The calculation formula is:

[0064]

[0065] therefore,

[0066] S8-2, because It is a collection of characteristic parameters. The representative parameters need to be obtained according to the following steps. The calculation method is as follows:

[0067] S8-2-1. Calculation Circumference and Perimeter ratio The calculation steps are:

[0068] perimeter The calculation formula is:

[0069]

[0070] at the same time, perimeter The calculation formula is:

[0071]

[0072] therefore,

[0073] S8-2-2. Calculation Area and Area ratio The calculation steps are:

[0074] area The calculation formula is:

[0075] Among them, A=2ax+by, B=2cy+bx;

[0076] area The calculation formula is:

[0077] therefore,

[0078] S8-2-3, and The integral of the curvature ratio The calculation steps are:

[0079] From the surface s The parameter expression is:

[0080] The first derivative of z:

[0081]

[0082] The second derivative of z:

[0083]

[0084] therefore, The curvature of (x, y, z) determined by any θ for: Where × is the vector cross product;

[0085] The curvature of (x, y) determined by any θ The calculation formula is:

[0086]

[0087] Where x = 0.5*cosθ*d min-x +(i+0.5)*d min-x , y=0.5*sinθ*d min-y +(j+0.5)*d min-y ;

[0088] so, The calculation formula is:

[0089]

[0090] The above technical solution is the core of this technical solution and also the most important innovation. Feature measurement is a key link in visual analysis. Some commonly used methods, such as edge detection, corner detection, feature point extraction, local binary mode, directional gradient, accelerated robustness and deep learning, usually have theoretical foundations involving knowledge such as image processing, signal processing, probability theory, statistics and linear algebra. These methods extract the features of the object to be identified from specific angles, which can reveal the visual commonalities of the object from some aspects, but rarely measure features from the perspective of the intrinsic geometric structure of the object, and lose a lot of important feature information.

[0091] The purpose of the present application is also to provide a tire bubble X-ray feature recognition device based on curved surface geometry, which is used to implement the above-mentioned tire bubble X-ray feature recognition method based on curved surface geometry, and includes the following parts:

[0092] X-ray machine photo data preprocessor, deployed on the edge, used to preprocess the acquired tire X-ray machine photos;

[0093] Data storage center, deployed in the cloud, for centralized storage of pre-processed X-ray machine photo data and subsequent calculation result data;

[0094] The data computing center, deployed in the cloud, introduces surface geometry analysis methods to obtain bubble characteristics from tire X-ray photo samples;

[0095] The bubble feature display and statistics system is used to visually display the bubbles identified from tire X-ray samples and the statistical data related to the bubble features;

[0096] Preferably, the X-ray machine photo data preprocessor is directly connected to the X-ray machine as an edge processing end for compressing photos output by the X-ray machine.

[0097] Preferably, the data storage center is composed of a server cluster, which is used to store compressed photo data obtained from an X-ray machine photo data preprocessor to form a data sample library, and is also used to store bubble feature data identified by a data computing center.

[0098] Preferably, the data computing center is deployed on a server cluster, reads tire X-ray photo data from the data storage center, and uses a bubble feature recognition algorithm to batch recognize bubble features from sample photos, and the recognition results are then stored in the data storage center.

[0099] Preferably, the bubble display and statistics system is deployed on a terminal computer for visually presenting bubble recognition results and statistical data thereof, and the terminal computer is connected to a data storage center via an Internet network.

[0100] The above technical scheme constitutes a complete set of tire bubble feature recognition devices, which can pre-process, store, calculate bubble features, display bubbles and perform feature statistical analysis on production site data in real time, effectively solving the difficulty of feature recognition caused by the diverse bubble shapes and blurred boundaries in traditional tire X-ray detection. Its advantage is that the surface geometry analysis principle is introduced in the bubble feature calculation link of the data computing center, and the bubble features are quantitatively described, thereby significantly enhancing the recognition ability of bubble features. It can be seen that it is a new technical solution for extracting bubble features under the conditions of tire X-ray machine photos. The obtained bubble feature results can not only be integrated with conventional defect analysis methods, but also with advanced intelligent recognition technologies, and have broad application prospects.

[0101] Compared with the prior art, the beneficial effects of this application are:

[0102] This application introduces surface geometry into feature measurement, which has the advantages of building a representation bridge between the intrinsic geometric structure of the object morphology and the object features, enhancing the feature measurement capability, and having wider adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] Figure 1 This is a schematic diagram of a tire bubble X-ray feature recognition device and method based on curved surface geometry disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0104] The present application is further described in detail below in conjunction with the accompanying drawings.

[0105] like Figure 1 As shown, a tire bubble X-ray feature recognition method based on curved surface geometry specifically includes the following steps:

[0106] S1. On the tire X-ray electronic photograph, mark the bubble area, take the upper left corner vertex as the coordinate origin, and define a three-dimensional vector (x, y, z) at each pixel point, where x represents the horizontal coordinate of the pixel, y represents the vertical coordinate of the pixel, and z represents the gray value of the pixel. The larger the z value, the higher the brightness of the corresponding pixel point on the picture. All (x, y, z) constitute a set P, the three-dimensional space where P is located is recorded as U, and the horizontal and vertical coordinates constitute the projection coordinate system V;

[0107] S2. In the projection coordinate system V, the distance between the centers of two adjacent pixels along the horizontal axis is d x , the distance along the ordinate axis is denoted as d y ;

[0108] S3, select a suitable positive integer C as an adjustable parameter;

[0109] S4, in the projection coordinate system V, with d min-y=C*d y Length is the interval, and several straight lines are drawn along the horizontal axis. min-x =C*d x Length is the interval, and several straight lines are drawn along the ordinate direction to grid the projection coordinate system V. Each grid Project P upward along the gray value to grid it and get several p i-j ,in, The four vertices are:

[0110]

[0111]

[0112]

[0113]

[0114] S5, fitting P to obtain the gray value surface s and its equation in the three-dimensional space;

[0115] The steps to calculate the surface s equation are as follows:

[0116] S5-1. Set a binary quadratic function to be determined: f(x, y) = ax 2 +bxy+cy 2 +d;

[0117] S5-2, the solution P has more than 4 points, that is, (x k ,y k , z k ), k=1, 2, 3,...N, N≥4;

[0118] S5-3, construct the following formula:

[0119] S5-4. Order:

[0120] S5-5. Solve the above four equations together to obtain a, b, c, d, which is the equation of the surface s: z = f (x, y) = ax 2 +bxy+cy 2 +d;

[0121] S6, any The rectangular area formed is recorded as The inscribed ellipse is denoted by

[0122] Calculate the inscribed ellipse The steps are as follows:

[0123] S6-1. Acquisition The four vertices of:

[0124]

[0125]

[0126]

[0127]

[0128] S6-2. Acquisition The center point:

[0129] S6-3, Upper inscribed ellipse The equation is:

[0130] S7, and Mapped onto the gray value surface s, we get the mapping structure and With p i-j correspond;

[0131] calculate and The steps are as follows:

[0132] S7-1. Obtain the equation of the surface s: z = f (x, y) = ax 2 +bxy+cy 2 +d;

[0133] S7-2, Calculation The steps are:

[0134] The equation of the projection of the first edge of on V is: y = j * d min-y , the two endpoints are: Substituting into the equation of surface s, we get The equation of the first side: z = ax 2 +b*j*d min-y *c+c*j 2 *d min-y 2 +d, where x∈[i*d min-x , (i+1)*d min-x ];

[0135] Using a similar principle, we can calculate The equations of the other three sides are:

[0136] z=a*(i+1)2 *d min-x 2 +b*(i+1)*d min-x *y+cy 2 +d, where y∈[j*d min-y , (j+1)*d min-y ];

[0137] z=ax 2 +b*(j+1)*d min-y *x+c*(j+1) 2 *d min-y 2 +d, where x∈[i*d min-x , (i+1)*d min-x ];

[0138] z=a*i 2 *d min-x 2 +b*i*d min-x *y+cy 2 +d, where y∈[j*d min-y , (j+1)*d min-y ];

[0139] S7-3. Calculation The steps are as follows:

[0140] Obtaining the inscribed ellipse based on S6-3 Parametric equation: x = 0.5*cosθ*d min-x +(i+0.5)*d min-x , y=0.5*sinθ*d min-y +(j+0.5)*d min-y ;

[0141] Substitute the above parametric equation into z = f (x, y) = ax 2 +bxy+cy 2 +d, we get the surface s The parametric equations are:

[0142] z=a(0.5*cosθ*d min-x +(i+0.5)*d min-x ) 2 +b*(0.5*cosθ*d min-x +(i+0.5)*d min-x )*(0.5*sinθ*d min-y +(j+0.5)*d min-y )+c(0.5*sinθ*d min-y +(j+0.5)*d min-y) 2 +d;

[0143] S8, calculate this time The characteristic parameter group And calculate The characteristic parameter group

[0144] calculate and The steps are as follows:

[0145] S8-1. Cause It is a collection of characteristic parameters. The representative parameters need to be obtained according to the following steps. The calculation method is as follows;

[0146] S8-1-1. Calculation Circumference and Perimeter ratio Here are the steps:

[0147] perimeter The calculation formula is:

[0148]

[0149] therefore,

[0150] S8-1-2. Calculation Area and Area ratio Here are the steps:

[0151] area The calculation formula is:

[0152]

[0153] therefore,

[0154] S8-1-3. Calculation sum of interior angles and Ratio of the sum of the interior angles Here are the steps:

[0155] sum of interior angles The calculation formula is:

[0156]

[0157] therefore,

[0158] S8-2, because It is a collection of characteristic parameters. The representative parameters need to be obtained according to the following steps. The calculation method is as follows:

[0159] S8-2-1. Calculation Circumference and Perimeter ratio The calculation steps are:

[0160] perimeter The calculation formula is:

[0161]

[0162] at the same time, perimeter The calculation formula is:

[0163]

[0164] therefore,

[0165] S8-2-2. Calculation Area and Area ratio The calculation steps are:

[0166] area The calculation formula is:

[0167] Among them, A=2ax+by, B=2cy+bx;

[0168] area The calculation formula is:

[0169] therefore,

[0170] S8-2-3, and The integral of the curvature ratio The calculation steps are:

[0171] From the surface s The parameter expression is:

[0172] The first derivative of z:

[0173]

[0174] The second derivative of z:

[0175]

[0176] therefore, The curvature of (x, y, z) determined by any θ for: Where × is the vector cross product;

[0177] The curvature of (x, y) determined by any θ The calculation formula is:

[0178]

[0179] Where x = 0.5*cosθ*d min-x +(i+0.5)*d min-x , y=0.5*sinθ*d min-y +(j+0.5)*d min-y ;

[0180] so, The calculation formula is:

[0181]

[0182] S9, obtain a large number of tire X-ray electronic photo bubble area samples, and for each type of bubble, calculate its and Get the bubble type and distributed;

[0183] S10, the bubble feature extraction calculation based on the tire X-ray photo is completed.

[0184] like Figure 1 As shown, the purpose of this application is also to provide a tire bubble X-ray feature recognition device based on curved surface geometry, comprising the following parts:

[0185] X-ray machine photo data preprocessor, deployed on the edge, used to preprocess the acquired tire X-ray machine photos;

[0186] Data storage center, deployed in the cloud, for centralized storage of pre-processed X-ray machine photo data and subsequent calculation result data;

[0187] The data computing center, deployed in the cloud, introduces surface geometry analysis methods to obtain bubble characteristics from tire X-ray photo samples;

[0188] The bubble feature display and statistics system is used to intuitively display the bubbles identified from tire X-ray samples and the statistical data related to the bubble features.

[0189] Preferably, the X-ray machine photo data preprocessor is directly connected to the X-ray machine as an edge processing end for compressing photos output by the X-ray machine.

[0190] Preferably, the data storage center is composed of a server cluster, which is used to store compressed photo data obtained from an X-ray machine photo data preprocessor to form a data sample library, and is also used to store bubble feature data identified by a data computing center.

[0191] Preferably, the data computing center is deployed on a server cluster, reads tire X-ray photo data from the data storage center, and uses a bubble feature recognition algorithm to batch recognize bubble features from sample photos, and the recognition results are then stored in the data storage center.

[0192] Preferably, the bubble display and statistics system is deployed on a terminal computer for visually presenting bubble recognition results and statistical data thereof, and the terminal computer is connected to a data storage center via an Internet network.

[0193] This embodiment aims to solve the problem of bubble recognition in tire X-ray images, and proposes a set of bubble feature recognition devices, including four parts: X-ray machine photo data preprocessor, data storage center, data calculation center, and bubble feature display and statistical system. The innovation of this embodiment is that, considering the diversity of bubbles in size, shape and distribution, a surface geometry method is creatively introduced to reveal the common characteristics of bubbles. Specifically, by analyzing the mapping deformation of two basic graphics on the plane - rectangle and ellipse - on the gray value surface, two groups of six feature indicators are designed. These indicators establish a characterization relationship between the bubble gray value surface geometry structure and the bubble characteristics, so that this embodiment has a wider recognition range and deeper feature analysis capabilities in identifying tire bubble characteristics, and has significant advantages. In addition, the device has the characteristics of simple deployment, and the corresponding calculation method is efficient and has good versatility. It is worth noting that the surface geometry method introduced in this embodiment is not only applicable to tire bubble recognition, but can also be extended to visual feature extraction problems in other fields, with high theoretical guidance and technical reliability.

Claims

1. A tire bubble X-ray feature recognition device based on curved surface geometry, characterized in that: It includes the following parts: X-ray machine photo data preprocessor, deployed on the edge, used to preprocess the acquired tire X-ray machine photos; Data storage center, deployed in the cloud, for centralized storage of pre-processed X-ray machine photo data and subsequent calculation result data; The data computing center, deployed in the cloud, introduces surface geometry analysis methods to obtain bubble characteristics from tire X-ray photo samples; The bubble feature display and statistics system is used to intuitively display the bubbles identified from tire X-ray samples and the statistical data related to the bubble features.

2. The tire bubble X-ray feature recognition device based on curved surface geometry according to claim 1, characterized in that: The X-ray machine photo data preprocessor is directly connected to the X-ray machine and serves as an edge processing end for compressing the photos output by the X-ray machine.

3. The tire bubble X-ray feature recognition device based on curved surface geometry according to claim 1, characterized in that: The data storage center is composed of a server cluster, which is used to store compressed photo data obtained from an X-ray machine photo data preprocessor to form a data sample library, and is also used to store bubble feature data identified by a data computing center.

4. The tire bubble X-ray feature recognition device based on curved surface geometry according to claim 1, characterized in that: The data computing center is deployed on a server cluster, reads tire X-ray photo data from the data storage center, and uses a bubble feature recognition algorithm to batch recognize bubble features from sample photos, and the recognition results are then stored in the data storage center.

5. The tire bubble X-ray feature recognition device based on curved surface geometry according to claim 1, characterized in that: The bubble display and statistics system is deployed on a terminal computer for visually presenting bubble recognition results and statistical data thereof. The terminal computer is connected to a data storage center via an Internet network.

6. A tire bubble X-ray feature recognition method based on curved surface geometry, implemented based on a tire bubble X-ray feature recognition device based on curved surface geometry as claimed in any one of claims 1 to 5, characterized in that: The following steps are involved: S1. On the tire X-ray electronic photograph, mark the bubble area, take the upper left corner vertex as the coordinate origin, and define a three-dimensional vector (x, y, z) at each pixel point, where x represents the horizontal coordinate of the pixel, y represents the vertical coordinate of the pixel, and z represents the gray value of the pixel. The larger the z value, the higher the brightness of the corresponding pixel point on the picture. All (x, y, z) constitute a set P, the three-dimensional space where P is located is recorded as U, and the horizontal and vertical coordinates constitute the projection coordinate system V; S2. In the projection coordinate system V, the distance between the centers of two adjacent pixels along the horizontal axis is d x , the distance along the ordinate axis is denoted as d y ; S3, select a suitable positive integer C as an adjustable parameter; S4, in the projection coordinate system V, with d min-y =C*d y Length is the interval, and several straight lines are drawn along the horizontal axis. min-x =C*d x Length is the interval, and several straight lines are drawn along the ordinate direction to grid the projection coordinate system V. Each grid Project P upward along the gray value to grid it and get several p i-j ,in, The four vertices are: S5, fitting P to obtain the gray value surface s and its equation in the three-dimensional space; S6, any The rectangular area formed is recorded as The inscribed ellipse is denoted by S7, and Mapped onto the gray value surface s, we get the mapping structure and With p i-j correspond; S8, calculate this time The characteristic parameter group And calculate The characteristic parameter group S9, obtain a large number of tire X-ray electronic photo bubble area samples, and for each type of bubble, calculate its and Get the bubble type and distributed; S10, the bubble feature extraction calculation based on the tire X-ray photo is completed.

7. The tire bubble X-ray feature recognition method based on curved surface geometry according to claim 6, characterized in that: In step S5, the steps of calculating the surface s equation are as follows: S5-1. Set a binary quadratic function to be determined: f(x, y) = ax 2 +bxy+cy 2 +d; S5-2, the solution P has more than 4 points, that is, (x k ,y k , z k ), k=1, 2, 3,...N, N≥4; S5-3, construct the following formula: S5-4. Order: S5-5. Solve the above four equations together to obtain a, b, c, d, which is the equation of the surface s: z = f(x, y) = ax 2 +bxy+cy 2 +d.

8. The tire bubble X-ray feature recognition method based on curved surface geometry according to claim 6, characterized in that: In step S6, Calculate the inscribed ellipse The steps are as follows: S6-1. Acquisition The four vertices of: S6-2. Acquisition Center point: S6-3, Upper inscribed ellipse The equation is:

9. The tire bubble X-ray feature recognition method based on curved surface geometry according to claim 7, characterized in that: In step S7, calculate and The steps are as follows: S7-1. Obtain the equation of the surface s: z = f (x, y) = ax 2 +bxy+cy 2 +d; S7-2, Calculation The steps are: The equation of the projection of the first edge of on V is: y = j * d min-y , the two endpoints are: Substituting into the equation of surface s, we get The equation of the first side of : Among them, x∈[i*d min-x , (i+1)*d min-x ]; Using a similar principle, we can calculate The equations of the other three sides are: z = a * (i + 1) 2 * d min-x 2 + b * (i + 1) * d min-x * y + cy 2 + d where y ∈ [j * d min-y , (j + 1) * d min-y ; z = ax 2 + b*(j + 1)*d min-y *x + c*(j + 1) 2 *d min-y 2 + d, where x ∈ [i*d min-x , (i + 1)*dm in-x ; z = a * i 2 * d min-x 2 + b * i * d min-x * y + cy 2 + d, where y ∈ [j * d min-y , (j + 1) * d min-y ; S7-3. Calculation The steps are as follows: Obtaining the inscribed ellipse based on S6-3 Parametric equation: x = 0.5*cosθ*d min-x +(i+0.5)*d min-x , y=0.5*sinθ*d min-y +(j+0.5)*d min-y ; Substitute the above parametric equation into z = f (x, y) = ax 2 +bxy+cy 2 +d, we get the surface s The parametric equations are: z=a(0.5*cos*d min-x +(i+0.5)*d min-x ) 2 +b*(0.5*cosθ*d min-x +(i+0.5)*d min-x )*(0.5*sinθ*d min-y +(j+0.5)*d min-y )+c(0.5*sinθ*d min-y +(j+0.5)*d min-y ) 2 +d。 10. The tire bubble X-ray feature recognition method based on curved surface geometry according to claim 6, characterized in that: In step S8, calculate and The steps are as follows: S8-1. Cause It is a collection of characteristic parameters. The representative parameters need to be obtained according to the following steps. The calculation method is as follows: S8-1-1. Calculation Circumference and Perimeter ratio Here are the steps: perimeter The calculation formula is: therefore, S8-1-2. Calculation Area and Area ratio Here are the steps: area The calculation formula is: therefore, S8-1-3. Calculation sum of interior angles and Ratio of the sum of the interior angles Here are the steps: sum of interior angles The calculation formula is: therefore, S8-2, because It is a collection of characteristic parameters. The representative parameters need to be obtained according to the following steps. The calculation method is as follows: S8-2-1. Calculation Circumference and Perimeter ratio The calculation steps are: perimeter The calculation formula is: at the same time, perimeter The calculation formula is: therefore, S8-2-2. Calculation Area and Area ratio The calculation steps are: area The calculation formula is: Among them, A=2ax+by, B=2cy+bx; area The calculation formula is: therefore, S8-2-3, and The integral of the curvature ratio The calculation steps are: From the surface s The parameter expression is: The first derivative of z: The second derivative of z: therefore, The curvature of (x, y, z) determined by any θ for: Where × is the vector cross product; The curvature of (x, y) determined by any θ The calculation formula is: where x = 0.5 * cosθ * d min-x + (i + 0.5) * d min-x , y = 0.5 * sinθ * d min-y + (j + 0.5) * d min-y ; so, The calculation formula is:

Images