A method and system for testing the durability of pantograph contact grids for highway trucks

Abstract

The present invention discloses a method and system for testing the durability of pantograph contact networks for highway trucks. The method comprises: step S1, using a high-definition area array camera to capture a pantograph image, and employing a Gabor filter and a support vector machine to identify a pantograph image to be tested in the pantograph image; step S2, using an improved edge detection algorithm to extract a pantograph edge image from the pantograph image to be tested; step S3, locating the pantograph contour based on the pantograph edge image, and obtaining pantograph contour similarity using an invariant moment and a similarity calculation formula; and step S4, obtaining the degree of wear of the pantograph contact network for highway trucks based on the contour similarity, and determining the durability of the pantograph contact network based on the wear degree. By employing a Gabor filter and a support vector machine (SVM) to identify the pantograph image, the present invention can effectively extract pantograph shape information and improve the recognition accuracy of the pantograph image to be tested.

Description

Technical Field

[0001] The present invention belongs to the technical field of pantograph detection, and in particular relates to a method and system for testing the durability of a pantograph contact grid for a highway truck. Background Art

[0002] Durability testing of pantographs and catenary networks for highway trucks presents a key technical challenge. During high-speed driving, the dynamic contact force between the pantograph and the catenary network fluctuates dramatically. This instability in contact force can lead to excessive wear or even damage to the pantograph's carbon plate. Furthermore, contact force fluctuations can also cause arcing between the pantograph and the catenary network. Arcing not only damages the pantograph and catenary network but can also impact the stability of the power supply system.

[0003] Furthermore, under complex and changing road and weather conditions, the pantograph's dynamic height can fluctuate frequently. If the pantograph's height control isn't precise and stable enough, it can degrade contact with the catenary grid, or even cause it to lose contact. This not only affects power supply continuity but can also cause mechanical damage to the pantograph and catenary grid. Summary of the Invention

[0004] To solve the above problems, the present invention proposes a method for testing the durability of a pantograph contact grid for highway trucks, the method comprising:

[0005] Step S1, using a high-definition area array camera to capture a pantograph image, and using a Gabor filter and a support vector machine to identify a pantograph image to be detected in the pantograph image;

[0006] Step S2, using the improved edge detection algorithm to extract the pantograph edge image from the pantograph image to be detected;

[0007] Step S3: locating the pantograph contour based on the pantograph edge image, and obtaining the pantograph contour similarity using a similarity calculation formula using an invariant moment;

[0008] Step S4: obtaining the degree of wear of the pantograph contact network for highway trucks based on the contour similarity, and determining the durability of the pantograph contact network based on the degree of wear.

[0009] Optionally, in step S1, the process of using a Gabor filter and a support vector machine to identify the pantograph image to be detected in the pantograph image specifically includes:

[0010] The pantograph image is filtered using a multi-scale directional Gobar filter to obtain pantograph shape information;

[0011] Convolution is performed on the shape information of the pantograph to extract a pantograph feature map, the pantograph feature map is divided into a plurality of grids, and the mean and variance of each grid are calculated;

[0012] The calculated mean and variance are serially connected into a feature descriptor, and a support vector machine is used to judge the pantograph based on the feature descriptor.

[0013] Optionally, the calculation formula for obtaining the pantograph shape information after filtering the pantograph image using a multi-scale directional Gobar filter is:

[0014] g(x,y)=s(x,y)×w(x,y)

[0015] Among them, s(x,y) is the sine function, w(x,y) is the Gaussian kernel function, and x and y are the horizontal and vertical pixel coordinates of the pantograph image.

[0016] Optionally, the Gaussian kernel function is:

[0017]

[0018] Optionally, in step S2, the improved edge detection algorithm is specifically:

[0019] Perform image denoising on the pantograph image to be detected using a self-median filter and a guided filter to obtain a denoised image;

[0020] Calculating pixel partial derivatives using the Sobel operator, and convolving the obtained pixel partial derivatives with the denoised image to obtain image gradient information;

[0021] The image gradient information is compared with two adjacent interpolation values ​​with the same gradient to obtain selected edge points, and the edge of the pantograph image to be detected is obtained based on the selected edge points.

[0022] Optionally, the calculation formula of the guided filter is:

[0023]

[0024] Among them, a k 、b k For the current window w k When the center of is located at k, the constant coefficient of the linear function, S is the guide image value, and i and k are pixel indexes.

[0025] Optionally, in step S3, the process of locating the pantograph contour based on the pantograph edge image and obtaining the pantograph contour similarity using the invariant moment and the similarity calculation formula specifically includes:

[0026] Calculating a geometric moment expression of the pantograph edge image in a discrete state with an order of p+q;

[0027] Calculating a central moment expression based on the geometric moment expression, and calculating a center distance normalization formula based on the center distance expression;

[0028] Based on the normalization formula, the pantograph profile similarity is calculated using the invariant moment through the similarity formula.

[0029] Optionally, the geometric moment expression is:

[0030]

[0031] Wherein, M and N are the maximum values ​​of the horizontal and vertical coordinates of the pantograph image to be detected.

[0032] The present invention also discloses a pantograph contact grid durability testing system for highway trucks, the system comprising:

[0033] An image acquisition module is used to acquire a pantograph image using a high-definition area array camera, and identify a pantograph image to be detected in the pantograph image using a Gabor filter and a support vector machine;

[0034] An edge acquisition module is used to extract a pantograph edge image from the pantograph image to be detected using an improved edge detection algorithm;

[0035] A similarity calculation module is used to locate the outline of the pantograph based on the pantograph edge image, and obtain the pantograph outline similarity by using an invariant moment and a similarity calculation formula;

[0036] The durability determination module is used to obtain the degree of wear of the pantograph contact network for highway trucks based on the contour similarity, and to determine the durability of the pantograph contact network based on the wear degree.

[0037] Optionally, in the image acquisition module, the process of using a Gabor filter and a support vector machine to identify the pantograph image to be detected in the pantograph image specifically includes:

[0038] The pantograph image is filtered using a multi-scale directional Gobar filter to obtain pantograph shape information;

[0039] Convolution is performed on the shape information of the pantograph to extract a pantograph feature map, the pantograph feature map is divided into a plurality of grids, and the mean and variance of each grid are calculated;

[0040] The calculated mean and variance are serially connected into a feature descriptor, and a support vector machine is used to judge the pantograph based on the feature descriptor.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] The present invention uses Gabor filters and support vector machines (SVM) to identify pantograph images, which can effectively extract the shape information of the pantograph and improve the recognition accuracy of the pantograph image to be detected. The improved edge detection algorithm is combined with the median filter and the guide filter for image denoising, which enhances the robustness of the system under different lighting and noise conditions. The pixel partial derivatives are calculated using the Sobel operator, and combined with the image gradient information, the edge image of the pantograph can be accurately extracted, providing accurate basic data for subsequent contour positioning. The contour of the pantograph is located based on the pantograph edge image, and the pantograph contour similarity is obtained through the similarity calculation formula using the invariant moment, which can accurately evaluate the degree of wear of the pantograph contact grid. Based on the evaluation of the degree of wear, the durability of the pantograph contact grid can be effectively judged, providing a scientific basis for maintenance and replacement. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0044] Figure 1 This is a method step diagram for testing the durability of a pantograph contact grid for highway trucks according to an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application.

[0046] First, some specific terms in this embodiment are explained and illustrated:

[0047] Pantograph, pantograph is an electrical device that electric traction locomotive obtains electricity from the contact network and is installed on the roof of the locomotive or EMU.

[0048] Area array camera: The target surface is rectangular or square. This embodiment uses several area array cameras to shoot the pantograph, so that the image has a stronger three-dimensional effect and less distortion.

[0049] Example 1

[0050] like Figure 1 As shown, this embodiment provides a method for testing the durability of a pantograph contact grid for a highway truck, the method comprising:

[0051] Step S1, using a high-definition area array camera to capture a pantograph image, and using a Gabor filter and a support vector machine to identify the pantograph image to be detected in the pantograph image, specifically including: using a multi-scale Gobar filter to filter the pantograph image to obtain the shape information of the pantograph; convolving the pantograph shape information to extract a pantograph feature map, dividing the pantograph feature map into several grids, and calculating the mean and variance of each grid; concatenating the calculated mean and variance into a feature descriptor, and using a support vector machine to judge the pantograph based on the feature descriptor.

[0052] The calculation formula for the pantograph shape information obtained by filtering the pantograph image using a multi-scale Gobar filter is:

[0053] g(x,y)=s(x,y)×w(x,y)

[0054] Among them, s(x,y) is the sine function, w(x,y) is the Gaussian kernel function, and x and y are the horizontal and vertical pixel coordinates of the pantograph image.

[0055] The sine function is:

[0056] s(x,y)=exp(i2πf0(xcosθ+ysinθ)).

[0057] The Gaussian kernel function is:

[0058]

[0059] The pantograph image (in this embodiment, because the area array camera is installed above the pantograph, the obtained image is a linear array image of the roof pantograph) is convolved with the pantograph shape information g(x, y) to obtain the pantograph feature map:

[0060] F(x,y)=I(x,y)×g(x,y).

[0061] The obtained feature map is divided into three types of grids: 1×1, 2×2, and 4×4. The mean and variance in each grid are calculated respectively. The calculated mean and variance are connected in series to form a feature descriptor, and the support vector machine is used to judge the pantograph based on the feature descriptor.

[0062] Step S2: using the improved edge detection algorithm to extract the pantograph edge image from the pantograph image to be detected.

[0063] The improved edge detection algorithm is specifically as follows: using a self-median filter and a guided filter to perform image denoising on the pantograph image to be detected to obtain a denoised image; using the Sobel operator to calculate the pixel partial derivative, and convolving the obtained pixel partial derivative with the denoised image to obtain image gradient information;

[0064] The calculation formula of the guided filter is:

[0065]

[0066] Among them, a k 、b k For the current window w k When the center of is located at k, the constant coefficient of the linear function, S is the guide image value, and i and k are pixel indexes.

[0067] By minimizing the difference between the output image and the input image, the value of the cost function is minimized and the constant coefficient a is obtained. k 、b k :

[0068]

[0069] Where t is the input image and ε is the penalty parameter, which is calculated using the least squares method:

[0070]

[0071] Among them, u k 、 is the filter window w k The mean and variance of all pixels in the filter window, |w| is the total number of pixels in the filter window.

[0072] The image gradient information is compared with two adjacent interpolation values ​​with the same gradient to obtain selected edge points, and the edge of the pantograph image to be detected is obtained based on the selected edge points.

[0073] This embodiment uses the Sobel operator in four directions to calculate the partial derivatives:

[0074]

[0075]

[0076] Among them, P x (i,j),P y (i,j),P 45 (i,j) and P 135 (i, j) is the partial derivative of pixel point (i, j) in four directions.

[0077] Calculate the gradient amplitude based on the above partial derivatives and the pantograph edge image:

[0078]

[0079] Step S3, locating the pantograph contour based on the pantograph edge image, and obtaining the pantograph contour similarity using an invariant moment and a similarity calculation formula, specifically comprising:

[0080] Calculate the geometric moment expression of the pantograph edge image in a discrete state when the order is p+q; calculate the central moment expression based on the geometric moment expression, and calculate the center distance normalization formula based on the center distance expression; based on the normalization formula, use the invariant moment to calculate the pantograph contour similarity through the similarity formula.

[0081] The geometric moment expression is:

[0082]

[0083] Wherein, M and N are the maximum values ​​of the horizontal and vertical coordinates of the pantograph image to be detected.

[0084] The central moment expression is:

[0085]

[0086] The pantograph profile similarity is calculated as:

[0087]

[0088] Step S4: obtaining the degree of wear of the pantograph contact network for highway trucks based on the contour similarity, and determining the durability of the pantograph contact network based on the degree of wear.

[0089] The degree of wear is represented based on the similarity result, and the smaller the similarity result, the smaller the degree of wear.

[0090] Example 2

[0091] A pantograph contact grid durability test system for highway trucks, the system comprising:

[0092] The image acquisition module is used to use a high-definition area array camera to capture pantograph images, and adopt a Gabor filter and a support vector machine to identify the pantograph image to be detected in the pantograph image. Specifically, the module comprises: using a multi-scale Gobar filter to filter the pantograph image to obtain the shape information of the pantograph; performing convolution on the pantograph shape information to extract a pantograph feature map, dividing the pantograph feature map into several grids, and calculating the mean and variance of each grid; concatenating the calculated mean and variance into a feature descriptor, and using a support vector machine to determine the pantograph based on the feature descriptor.

[0093] The calculation formula for the pantograph shape information obtained by filtering the pantograph image using a multi-scale Gobar filter is:

[0094] g(x,y)=s(x,y)×w(x,y)

[0095] Among them, s(x,y) is the sine function, w(x,y) is the Gaussian kernel function, and x and y are the horizontal and vertical pixel coordinates of the pantograph image.

[0096] The sine function is:

[0097] s(x,y)=exp(i2πf0(xcosθ+ysinθ)).

[0098] The Gaussian kernel function is:

[0099]

[0100] The pantograph image (in this embodiment, because the area array camera is installed above the pantograph, the obtained image is a linear array image of the roof pantograph) is convolved with the pantograph shape information g(x, y) to obtain the pantograph feature map:

[0101] F(x,y)=I(x,y)×g(x,y).

[0102] The obtained feature map is divided into three types of grids: 1×1, 2×2, and 4×4. The mean and variance in each grid are calculated respectively. The calculated mean and variance are connected in series to form a feature descriptor, and the support vector machine is used to judge the pantograph based on the feature descriptor.

[0103] The edge acquisition module is used to extract the pantograph edge image from the pantograph image to be detected using the improved edge detection algorithm.

[0104] The improved edge detection algorithm is specifically as follows: using a self-median filter and a guided filter to perform image denoising on the pantograph image to be detected to obtain a denoised image; using the Sobel operator to calculate the pixel partial derivative, and convolving the obtained pixel partial derivative with the denoised image to obtain image gradient information;

[0105] The calculation formula of the guided filter is:

[0106]

[0107] Among them, a k 、b k For the current window w k When the center of is located at k, the constant coefficient of the linear function, S is the guide image value, and i and k are pixel indexes.

[0108] By minimizing the difference between the output image and the input image, the value of the cost function is minimized and the constant coefficient a is obtained. k 、b k :

[0109]

[0110] Where t is the input image and ε is the penalty parameter, which is calculated using the least squares method:

[0111]

[0112] Among them, u k 、 is the filter window w k The mean and variance of all pixels in the filter window, |w| is the total number of pixels in the filter window.

[0113] The image gradient information is compared with two adjacent interpolation values ​​with the same gradient to obtain selected edge points, and the edge of the pantograph image to be detected is obtained based on the selected edge points.

[0114] This embodiment uses the Sobel operator in four directions to calculate the partial derivatives:

[0115]

[0116] Among them, P x (i,j),P y (i,j),P 45 (i,j) and P 135 (i, j) is the partial derivative of pixel point (i, j) in four directions.

[0117] Calculate the gradient amplitude based on the above partial derivatives and the pantograph edge image:

[0118]

[0119] A similarity calculation module is used to locate the outline of the pantograph based on the pantograph edge image and obtain the pantograph outline similarity using an invariant moment and a similarity calculation formula, specifically including:

[0120] Calculate the geometric moment expression of the pantograph edge image in a discrete state when the order is p+q; calculate the central moment expression based on the geometric moment expression, and calculate the center distance normalization formula based on the center distance expression; based on the normalization formula, use the invariant moment to calculate the pantograph contour similarity through the similarity formula.

[0121] The geometric moment expression is:

[0122]

[0123] Wherein, M and N are the maximum values ​​of the horizontal and vertical coordinates of the pantograph image to be detected.

[0124] The central moment expression is:

[0125]

[0126] The pantograph profile similarity is calculated as:

[0127]

[0128] The durability determination module is configured to determine the degree of wear of the pantograph contact network for highway trucks based on the profile similarity, and to determine the durability of the pantograph contact network based on the wear degree. The wear degree is represented by the similarity result, and a smaller similarity result indicates a smaller wear degree.

[0129] The embodiments of the present disclosure are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present disclosure should be included in the scope of protection of the present disclosure.

Claims

1. A method for testing the durability of a pantograph contact grid for a highway truck, characterized in that: The method specifically includes: Step S1, using a high-definition area array camera to capture a pantograph image, and using a Gabor filter and a support vector machine to identify a pantograph image to be detected in the pantograph image; Step S2, using the improved edge detection algorithm to extract the pantograph edge image from the pantograph image to be detected; In step S2, the improved edge detection algorithm is specifically as follows: Perform image denoising on the pantograph image to be detected using a self-median filter and a guided filter to obtain a denoised image; Calculating pixel partial derivatives using the Sobel operator, and convolving the obtained pixel partial derivatives with the denoised image to obtain image gradient information; Comparing the image gradient information with two adjacent interpolation values ​​of the same gradient to obtain selected edge points, and obtaining the edge of the pantograph image to be detected based on the selected edge points; The calculation formula of the guided filter is: Among them, a k 、b k For the current window w k When the center of is located at k, the constant coefficient of the linear function, S is the guide image value, i and k are pixel indexes; Step S3: locating the pantograph contour based on the pantograph edge image, and obtaining the pantograph contour similarity using a similarity calculation formula using an invariant moment; Step S4: obtaining the degree of wear of the pantograph contact network for highway trucks based on the contour similarity, and determining the durability of the pantograph contact network based on the degree of wear.

2. The method for testing the durability of a pantograph contact grid for a highway truck according to claim 1, characterized in that: In step S1, the process of using a Gabor filter and a support vector machine to identify the pantograph image to be detected in the pantograph image specifically includes: The pantograph image is filtered using a multi-scale directional Gobar filter to obtain pantograph shape information; Convolution is performed on the shape information of the pantograph to extract a pantograph feature map, the pantograph feature map is divided into a plurality of grids, and the mean and variance of each grid are calculated; The calculated mean and variance are serially connected into a feature descriptor, and a support vector machine is used to judge the pantograph based on the feature descriptor.

3. The method for testing the durability of a pantograph contact grid for a highway truck according to claim 2, characterized in that: The calculation formula for the pantograph shape information obtained by filtering the pantograph image using a multi-scale Gobar filter is: g(x,y)=s(x,y)×w(x,y) Among them, s(x,y) is the sine function, w(x,y) is the Gaussian kernel function, and x and y are the horizontal and vertical pixel coordinates of the pantograph image.

4. The method for testing the durability of a pantograph contact grid for a highway truck according to claim 3, characterized in that: The Gaussian kernel function is:

5. The method for testing the durability of a pantograph contact grid for a highway truck according to claim 1, characterized in that: In step S3, the process of locating the pantograph contour based on the pantograph edge image and obtaining the pantograph contour similarity using the invariant moment and the similarity calculation formula specifically includes: Calculate the pantograph edge image in a discrete state, the order is p+q The geometric moment expression when ; Calculating a central moment expression based on the geometric moment expression, and calculating a center distance normalization formula based on the center distance expression; Based on the normalization formula, the pantograph profile similarity is calculated using the invariant moment through the similarity formula.

6. The method for testing the durability of a pantograph contact grid for a highway truck according to claim 5, characterized in that: The geometric moment expression is: Wherein, M and N are the maximum values ​​of the horizontal and vertical coordinates of the pantograph image to be detected.

7. A pantograph contact grid durability testing system for highway trucks, the system applying the method according to any one of claims 1 to 6, characterized in that: The system includes: An image acquisition module is used to acquire a pantograph image using a high-definition area array camera, and identify a pantograph image to be detected in the pantograph image using a Gabor filter and a support vector machine; An edge acquisition module is used to extract a pantograph edge image from the pantograph image to be detected using an improved edge detection algorithm; A similarity calculation module is used to locate the outline of the pantograph based on the pantograph edge image, and obtain the pantograph outline similarity by using an invariant moment and a similarity calculation formula; The durability determination module is used to obtain the degree of wear of the pantograph contact network for highway trucks based on the contour similarity, and to determine the durability of the pantograph contact network based on the wear degree.

8. The pantograph contact grid durability testing system for highway trucks according to claim 7, characterized in that: In the image acquisition module, the process of using the Gabor filter and the support vector machine to identify the pantograph image to be detected in the pantograph image specifically includes: The pantograph image is filtered using a multi-scale directional Gobar filter to obtain pantograph shape information; Convolution is performed on the shape information of the pantograph to extract a pantograph feature map, the pantograph feature map is divided into a plurality of grids, and the mean and variance of each grid are calculated; The calculated mean and variance are serially connected into a feature descriptor, and a support vector machine is used to judge the pantograph based on the feature descriptor.

Images