Bullet train bottom axle positioning method based on image gradient detection

Through image gradient detection and polynomial fitting, the center position of the bottom axle body of the EMU is automatically identified, solving the problem of low positioning efficiency of the bottom axle in the existing technology, and achieving more efficient and accurate positioning and identification.

CN118967778APending Publication Date: 2024-11-15HARBIN KEJIA GENERAL MECHANICAL & ELECTRICAL CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202410966083.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The existing EMU bottom axle positioning method is inefficient and depends on numerical fitting. It takes a long time and is not efficient enough, making it difficult to quickly and accurately obtain the bottom axle position.

Method used

Using an image gradient detection method, by obtaining the vehicle bottom height data, edge detection and center screening are performed, and combined with polynomial fitting, the center position of the vehicle bottom axle body is automatically identified to reduce manual intervention and noise impact.

Benefits of technology

It improves the efficiency and accuracy of the vehicle bottom axle positioning, shortens the recognition time, enhances the consistency of the recognition results, effectively eliminates noise points, and improves the recognition accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118967778A_ABST
    Figure CN118967778A_ABST
Patent Text Reader

Abstract

The invention discloses a bullet train bottom axle positioning method based on image gradient detection, solves the problem of how to improve the train bottom axle positioning efficiency, and belongs to the technical field of bullet train bottom detection. The method comprises the following steps: acquiring height data of each position obtained by scanning the bottom of the bullet train by a train bottom detection robot; drawing the height data of each position in a two-dimensional rectangular coordinate system, and taking an obtained drawn curve as a cross section profile diagram of the bottom of the bullet train; carrying out edge detection on the cross section contour diagram of the bottom of the bullet train to obtain a binarized image I; searching a circle in the binarized image I by using a gradient method to obtain a candidate circle center; all the obtained candidate circle centers are screened, and the center position coordinates of the shaft body at the bottom of the bullet train are obtained according to the remaining circle centers after screening. The method based on image gradient calculation is used for target searching, compared with a traditional numerical value fitting method, the time is shorter, the efficiency is higher, human intervention is not needed, and the manual workload is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a motor vehicle bottom axle positioning method based on image gradient detection, belonging to the technical field of motor vehicle bottom detection. Background Art

[0002] During its operation, the undercarriage inspection robot (RGV) needs to know its position relative to the vehicle being inspected accurately, so that it can accurately find the target operation point for inspection and avoid collision with the vehicle being inspected. This requires the inspection robot to have a corresponding positioning system and algorithm. The inspection robot is equipped with a vertically upward laser rangefinder on its body, which can obtain the precise height of each position under the vehicle. The existing method of undercarriage axle positioning is to obtain the cross-sectional profile of the undercarriage based on the data scanned by the inspection robot, and then use the circle drawn according to each axle as the standard, and use the numerical fitting method to find the position of each axle in the cross-sectional profile of the undercarriage. This method is time-consuming and inefficient. Summary of the invention

[0003] In order to improve the efficiency of vehicle bottom axle positioning, the present invention provides a vehicle bottom axle positioning method based on image gradient detection.

[0004] A method for positioning a motor vehicle bottom axle based on image gradient detection of the present invention comprises:

[0005] S1. Obtain the height data of each position obtained by the vehicle bottom inspection robot scanning the vehicle bottom:

[0006] S2, plotting the height data of each position in a two-dimensional rectangular coordinate system, and obtaining a plotted curve as a cross-sectional profile of the bottom of the train;

[0007] S3, performing edge detection on the cross-sectional profile of the bottom of the train to obtain a binary image I;

[0008] S4, using the gradient method to find the circle in the binary image I and obtain the candidate center of the circle;

[0009] S5. Screen all the candidate circle centers obtained, and obtain the coordinates of the axle center position of the bottom of the train according to the circle centers remaining after the screening.

[0010] Preferably, a gradient method is used to find a circle in the binary image I to obtain a candidate circle center:

[0011] Get the gradient direction of all pixels in the binary image I, traverse all non-zero pixels in the binary image I, draw a straight line along the gradient direction of each non-zero pixel, and for each non-zero pixel (x n ,y n )Set up an accumulator S(x n,y n ), used to record the crossing point (x n ,y n ) number of straight lines;

[0012] Set the threshold K, all the conditions that meet s(x n ,y n )>K point (x n ,y n ) is the candidate center.

[0013] As a preferred method, the gradient direction of all pixels in the binary image I is obtained as follows:

[0014] Perform Sobel operator convolution operations on the binary image I in the x direction and y direction respectively to obtain the neighborhood gradient values ​​of all pixels in the two directions, which are G and x and G y ;

[0015] Use the arctan function to find the gradient direction θ(x,y) of all pixels:

[0016] θ(x,y)=arctan(G y (x,y),G x (x,y)).

[0017] As a preference,

[0018] Where I is the identity matrix.

[0019] As a preferred method, all candidate circle centers are screened:

[0020] According to the known height and radius of the axle body under the EMU, the candidate circle centers that do not match are filtered out.

[0021] As a preferred method, according to the remaining circle centers after screening, the method of obtaining the coordinates of the axle center position of the motor vehicle bottom is as follows:

[0022] For each center coordinate (x0, y0) left after screening, find the point on the drawn curve corresponding to the horizontal coordinate x0 of the center in S2, extend the point to the left and right sides of the point by a set distance d, and intercept the [x0-d, x0+d] segment on the drawn curve;

[0023] Use a third-order polynomial to fit the [x0-d, x0+d] segment and get the analytical expression:

[0024] y=ax 3 +bx 2 +cx+d

[0025] Among them, a, b, c, and d are analytical parameters fitted;

[0026] Find the minimum point P of the analytical expression on [x0-d,x0+d] min =(x min ,y min ), where x min is the axis center coordinate.

[0027] Preferably, S1 further includes filtering the height data of each position.

[0028] The present invention also provides a computer-readable storage device, which stores a computer program, characterized in that when the computer program is executed by a processor, the steps of the above-mentioned method for locating the bottom axle of a motor vehicle based on image gradient detection are implemented.

[0029] The present invention also provides a motor vehicle underbody axle positioning device based on image gradient detection, comprising a storage device, a processor, and a computer program stored in the storage device and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-mentioned motor vehicle underbody axle positioning method based on image gradient detection.

[0030] The present invention also provides a computer program product, including a computer program, characterized in that when the computer program is executed by a processor, the steps of the above-mentioned method for locating the bottom axle of a motor vehicle based on image gradient detection are implemented.

[0031] The beneficial effects of the present invention are as follows: the present invention automatically obtains the axle position according to the data scanned by the detection robot, without human intervention, thus reducing the manual workload. The present invention uses a method based on image gradient calculation for target search, which is shorter and more efficient than the traditional numerical fitting method, and does not strictly require the data form. The present invention uses a polynomial fitting method to further improve the accuracy of the results and improve the consistency of multiple runs in practical applications. The present invention uses a filtering method to effectively eliminate noise points in the original data and improve recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a schematic diagram of the vehicle bottom scanning;

[0033] Figure 2 This is the scanned image of the vehicle bottom after removing the noise;

[0034] Figure 3 It is the suspected axis found based on the gradient detection method;

[0035] Figure 4 The coordinates of the axis center obtained by polynomial fitting are marked by two arrows in the figure;

[0036] Figure 5 This is a flow chart of the EMU vehicle bottom axle positioning method based on gradient detection. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0039] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0040] The method for locating the bottom axle of a motor vehicle based on image gradient detection of this embodiment includes:

[0041] Step 1: Obtain height data of each position obtained by scanning the bottom of the train by the bottom inspection robot;

[0042] Specifically, the processing object of this embodiment is the scanning result of the laser rangefinder on the body of the underbody inspection robot. The rangefinder is installed on the front of the fuselage, and the laser beam is vertically upward. The robot can obtain the height of the current laser beam irradiation point in real time. During sampling, the inspection robot turns on the laser beam and walks along the body of the vehicle under the inspected vehicle. During the walking process, the laser beam is shot at the underbody structure of the inspected vehicle, and the inspection robot saves the collected height data of each position on the disk. The saved height data is a set of binary coordinate points P = [(x1, y1), (x2, y2),…, x N ,y N )],x n Indicates the horizontal coordinate of the vehicle bottom detection robot, y n Represents the horizontal coordinate x n The table shows the height of the scanned vehicle bottom structure, n = 1, 2, ..., N, where N represents the number of scanning points.

[0043] Since the robot itself is in motion during the acquisition process, coupled with the influence of ambient lighting and debris under the vehicle, there may be noise points in the data returned by the laser rangefinder. Since the subsequent recognition process is based on graphic data, after obtaining the original acquired data, it is necessary to filter the data to eliminate the noise as much as possible. The filtering method can be limited filtering or median filtering.

[0044] Step 2, plotting the height data of each position in a two-dimensional rectangular coordinate system, and obtaining a plotted curve as a cross-sectional profile of the bottom of the train;

[0045] Specifically, for the filtered coordinate point data, the data is first adjusted to unify the length measurement units in the x and y directions, and then plotted in a two-dimensional rectangular coordinate system. The resulting curve can be approximately regarded as a cross-sectional profile of the bottom of the inspected vehicle.

[0046] Step 3: Perform edge detection on the cross-sectional profile of the bottom of the train to obtain a binary image I;

[0047] Specifically, the cross-sectional profile of the train bottom is processed by Canny edge detection to obtain a binarized image I. The purpose of this step is to further improve the sharpness of the image and reduce the influence of noise.

[0048] Step 4: Use the gradient method to find the circle in the binary image I and obtain the candidate center;

[0049] The specific steps are: perform Sobel operator convolution operations on the binary image I in the x direction and y direction respectively to obtain the neighborhood gradient values ​​of all pixels in the two directions, denoted as G x and G y ;

[0050]

[0051] Based on this, the inverse tangent function arctan is used to find the gradient direction of all pixels:

[0052] θ(x,y)=arctan(G y (x,y),G x (x,y)

[0053] Then traverse all non-zero pixels in the image and draw a straight line along the gradient direction of each non-zero pixel. n ,y n ), set an accumulator for it and record it as S(x n ,y n ). Traverse all non-zero pixels in the image and assign their corresponding S(x n ,y n ) is assigned a value of n, where n is the number of lines passing through this point. According to the characteristics of a circle, the gradient direction on the edge of its graph all points to the center of the circle. Therefore, the more lines pass through a point, the more likely it is that this point is the center of a circle. Set a threshold K, specifying that all lines that meet the condition S(x n ,y n )>K point (x n ,y n) are all candidate centers.

[0054] Step 5: Screen all the candidate circle centers obtained, and obtain the coordinates of the axle center position of the bottom of the train according to the remaining circle centers after screening;

[0055] For all the candidate centers, in addition to the real axle, there are some interference items that need to be eliminated. Since the height and radius of each EMU axle are known data, we only need to filter out those candidates whose height or radius does not match the real data. The remaining ones are the target points, and the horizontal coordinate of the center of the circle is the coordinate of the corresponding axle position.

[0056] However, due to the influence of foreign objects, vibration during the robot's walking process and other factors, the axis part of the original data curve may be deformed, resulting in it not being a standard arc, so the center of the circle may not be at the actual center of the axis. In order to eliminate these influences, it is necessary to further determine the coordinates of the axis center position of the EMU bottom;

[0057] For each center coordinate (x0, y0) left after screening, find the point on the drawn curve corresponding to the horizontal coordinate x0 of the center in S2, extend the point to the left and right sides of the point by a set distance d, and intercept the [x0-d, x0+d] segment on the drawn curve;

[0058] Use a third-order polynomial to fit the [x0-d, x0+d] segment and get the analytical expression:

[0059] y=ax 3 +bx 2 +cx+d

[0060] Among them, a, b, c, and d are analytical parameters fitted;

[0061] Since the cross section of the shaft is circular, the scanning curve corresponding to the position of the shaft center must be at the lowest point of the arc. Therefore, find the minimum point P of the analytical expression on [x0-d, x0+d] min =(x min ,y min ), where x min is the axis center coordinate.

[0062] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. It should therefore be understood that many modifications may be made to the exemplary embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in a manner different from that described in the original claims. It should also be understood that features described in conjunction with individual embodiments may be used in other described embodiments.

Claims

1. A method for locating the underbody axle of a motor vehicle based on image gradient detection, characterized in that: The method comprises: S1. Obtain the height data of each position obtained by the vehicle bottom inspection robot scanning the vehicle bottom: S2, plotting the height data of each position in a two-dimensional rectangular coordinate system, and obtaining a plotted curve as a cross-sectional profile of the bottom of the train; S3, performing edge detection on the cross-sectional profile of the bottom of the train to obtain a binary image I; S4, using the gradient method to find the circle in the binary image I and obtain the candidate center of the circle; S5. Screen all the candidate circle centers obtained, and obtain the coordinates of the axle center position of the bottom of the train according to the circle centers remaining after the screening.

2. The method for locating the underbody axle of a motor vehicle based on image gradient detection according to claim 1 is characterized in that: Use the gradient method to find the circle in the binary image I and get the candidate center: Get the gradient direction of all pixels in the binary image I, traverse all non-zero pixels in the binary image I, draw a straight line along the gradient direction of each non-zero pixel, and for each non-zero pixel (x n ,y n )Set up an accumulator S(x n ,y n ), used to record the crossing point (x n ,y n ) number of straight lines; Set the threshold K, all the conditions S(x n ,y n )>K's point (x n ,y n ) is the candidate center.

3. The method for locating the underbody axle of a motor vehicle based on image gradient detection according to claim 2 is characterized in that: The method to obtain the gradient direction of all pixels in the binary image I is: Perform Sobel operator convolution operations on the binary image I in the x direction and y direction respectively to obtain the neighborhood gradient values ​​of all pixels in the two directions, which are G and x and G y ; Use the arctan function to find the gradient direction θ(x,y) of all pixels: θ(x,y)=arctan(G y (x,y), G x (x,y))。 4. The method for locating the underbody axle of a motor vehicle based on image gradient detection according to claim 3 is characterized in that: Where I is the identity matrix.

5. The method for locating the underbody axle of a motor vehicle based on image gradient detection according to claim 1, characterized in that: Method for screening all candidate circle centers obtained: According to the known height and radius of the axle body under the EMU, the candidate circle centers that do not match are filtered out.

6. The method for locating the underbody axle of a motor vehicle based on image gradient detection according to claim 1, characterized in that: The method to obtain the coordinates of the axle center position of the bottom of the train according to the remaining circle centers after screening: For each center coordinate (x0, y0) left after screening, find the point on the drawn curve corresponding to the horizontal coordinate x0 of the center in S2, extend the point to the left and right sides of the point by a set distance d, and intercept the [x0-d, x0+d] segment on the drawn curve; Use a third-order polynomial to fit the [x0-d, x0+d] segment and get the analytical expression: y=ax 3 +bx 2 +cx+d Among them, a, b, c, and d are analytical parameters fitted; Find the minimum point P of the analytical expression on [x0-d,x0+d] min =(x min ,y min ), where x min is the axis center coordinate.

7. The method for locating the underbody axle of a motor vehicle based on image gradient detection according to claim 1, characterized in that: S1 also includes filtering the height data of each position.

8. A computer-readable storage device storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for locating the underbody axle of a motor vehicle based on image gradient detection as claimed in any one of claims 1 to 7 are implemented.

9. A motor vehicle underbody axle positioning device based on image gradient detection, comprising a storage device, a processor, and a computer program stored in the storage device and executable on the processor, characterized in that: The processor executes the computer program to implement the steps of the method for locating the underbody axle of a motor vehicle based on image gradient detection as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method for locating the underbody axle of a motor vehicle based on image gradient detection as described in any one of claims 1 to 7 are implemented.

Citation Information

Cited By

  • Tour inspection robot and electric bus traffic control method and system

    CN121341249A