CN110827239A - A method for measuring the lateral movement of train wheels - Google Patents

Abstract

The invention discloses a method for measuring the transverse movement of train wheels, which comprises the steps of arranging a binocular camera and a structured light source on a train bogie, calibrating the binocular camera to obtain internal and external parameters, enabling a plurality of pieces of structured light to irradiate on a wheel to be measured and a track at different angles, extracting the structured light in a preprocessed left image and a preprocessed right image through probability Hough linear detection, extracting end points at two ends of a linear segment of the structured light, performing three-dimensional reconstruction on a circle where the wheel is located and the linear segment of the track, calculating the transverse movement of the circle center of the wheel between a projection point of the plane of the track and the track, and judging whether abnormity exists. The invention can realize real-time monitoring of the transverse movement of the wheel rail in the running process of the train and guarantee the running safety.

Description

A method for measuring the lateral movement of train wheels

technical field

The invention relates to the technical field of measurement and analysis of the serpentine motion of a railway locomotive, in particular to a method for measuring the lateral movement of a train wheel.

Background technique

With the development of economy and society, the demand for transportation has soared. As the recognized means of transportation with the largest capacity, the market demand of railway locomotives is even tighter. Since the train turns through special structures such as wheel set treads and bogies, the serpentine motion of the train moving forward and swinging laterally during the straight running process is inevitable. Excessive serpentine motion will not only disrupt the smooth running of the vehicle, but also damage the wheels and rails, and may even lead to derailment and cause major accidents. Therefore, monitoring the serpentine movement of trains to ensure the safety of driving is very important for people's personal safety and property safety.

Therefore, the present invention hopes to measure the lateral movement of the train wheel in real time by designing a method for measuring the lateral movement of the train wheel, and when the lateral movement exceeds a certain range, the abnormal situation can be detected and notified in time.

SUMMARY OF THE INVENTION

The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, and proposes a method for measuring the lateral movement of train wheels based on structured light. Monitoring of train running safety in order to improve the safety of railway transportation.

In order to achieve the above purpose, the technical solution provided by the present invention is: a method for measuring the lateral movement of a train wheel, comprising the following steps:

1) A binocular camera and a structured light source are arranged on the train bogie. The shooting range includes the wheel and track to be tested. Adjust the structured light source and project the linear structured light on the wheel and track to be tested; System calibration, obtain the internal parameters of the binocular camera and perform stereo calibration to obtain the relative position relationship matrix between the binocular cameras, and use the internal and external parameters to perform stereo correction preprocessing on the left and right images;

2) Perform probabilistic Hough line detection on the image to extract the structured light lines in the left and right views; take the straight line angle as the matching feature, match the structured light segments in the left and right images one by one, and then match the same straight line segment The two end points of the two end points are matched; the binocular stereo vision algorithm is used to calculate the three-dimensional space coordinates of the feature points through the binocular internal parameters and external parameters obtained by calibration;

3) According to the spatial coordinates of the corresponding feature points, three-dimensional reconstruction of the wheel and the track is carried out, and the three-dimensional space equation of the wheel and the track is established;

4) Calculate the distance between the projection point of the wheel center on the track plane and the edge of the track as the lateral movement of the train wheel, and set the normal lateral movement range. If there is a distance between the projection point of the wheel center on the track plane and the edge of the track If the distance between the projection point of the wheel center on the track plane and the edge of the track exceeds the normal traverse range, it is determined that the train is running abnormally.

In step 1), a binocular camera and a plurality of structured light sources placed in parallel are used, placed on the train bogie, and the angle between the camera and the structured light source is adjusted, so that the binocular camera can capture the structured light irradiated at different angles. Images on wheels and tracks.

平移因子

旋转因子

表示令左摄像头与右摄像头具有相同姿态的旋转效果，具体能够拆分为沿x、y、z轴分别旋转θ_x、θ_y、θ_z这三个步骤，即旋转因子中的参数r₁～r₉具体由3个旋转矩阵的乘积决定：In step 1), the binocular camera is calibrated by Zhang Zhengyou's calibration method, and the internal parameters including S _X , S _Y , u ₀ , v ₀ , f are obtained, wherein S _X and S _Y are the physical dimensions of a single pixel of the camera chip , u ₀ , v ₀ are the pixel coordinates of the optical center of the optical axis, f is the focal length of the camera; perform stereo calibration on the relative positional relationship of the binocular camera, and obtain the external parameters R and T, where the rotation factor

translation factor

twiddle factor

Represents the rotation effect of making the left camera and the right camera have the same attitude, which can be divided into three steps of rotating θ _x , θ _y , and θ _z along the x, y, and z axes respectively, that is, the parameter r ₁ ~ r ₉ is specifically determined by the product of 3 rotation matrices:

t _x , _ty , and t _z in the translation factor are the translation distances of the origin of the right camera relative to the left camera. Let the coordinate system of the left camera be X _L and the coordinate system of the right camera be X _R , and the relative formula of the two cameras can be obtained as X _{R =} RXL+T.

In step 2), the probability Hough line detection is performed on the image, the straight line mapped by the structured light on the image is detected, the endpoint of the structured light on the edge of the object is selected as the feature point and matched, and after calculation, three non-collinear lines on the wheel are selected The feature points P ₁ (x ₁ , y ₁ , z ₁ ), P ₂ (x ₂ , y ₂ , z ₂ ), P ₃ (x ₂ , y ₂ , z ₂ ) and 2 feature points on the track side Q ₁ , Q ₂ and a feature point Q ₃ on the other side;

In step 3), the three non-collinear points are used to determine a plane, the wheel edge points P ₁ , P ₂ , P ₃ have equal distances from the center of the circle and are all R, and set the constants A _i , B _i , C _i , D _i , where (A _i ,B _i ,C _i ) is the normal vector of the wheel plane, and the solution center O(x _i ,y _i ,z _i ) space coordinates:

A _i x+B _i y+C _i z+D _i =0

其中(x₀,y₀,z₀)为直线l上的某点，(m,n,p)为直线l的方向向量；设定常数A、B、C、D，利用轨道同一侧的2个特征点Q₁、Q₂及另一侧的1个特征点Q₃，计算出车轮所在平面α的一般方程：Ax+By+Cz+D＝0，其中(A,B,C)为平面α的法向量；Set constants m, n and p, and use the two characteristic points Q ₁ and Q ₂ on the same side of the orbit to calculate the equation of the space straight line l where the orbit is located:

Where (x ₀ , y ₀ , z ₀ ) is a point on the straight line l, (m, n, p) is the direction vector of the straight line l; set the constants A, B, C, D, use the 2 on the same side of the track The general equation of the plane α where the wheel is located is calculated: Ax+By+Cz+D _{= 0} , where ( _A , B, C) is the plane the normal vector of α;

In step 4), calculate the projection _V (x,y, _z ) of the center of the circle on the orbital plane and the distance d between _V and the edge of the orbital; Intersecting at the projection point V(x, y, z), the parametric equation of the vertical line is obtained:

The parametric equation of the vertical line is brought into the orbit plane equation α, and the parameter t corresponding to the projection point V(x, y, z) in the parametric equation is obtained:

Then bring the parameter t into the parametric equation of the straight line to obtain the specific spatial coordinates V(x, y, z) of the projection point; use the two points Q ₁ and Q ₂ on the same side of the track edge to find the projection point V to the track edge the distance:

Set the normal traverse range [d _min , d _max ]. If the distance between the projection point of the circle center on the track plane and the track edge is d _min <d < d _max , the train runs normally; if the circle center is at the projection point of the track plane If the distance from the edge of the track exceeds the normal traverse range, the train runs abnormally at this time.

Compared with the prior art, the present invention has the following advantages and beneficial effects:

1. Real-time monitoring of the lateral movement formed between the wheel and the track during the driving process ensures the driving safety.

2. Through the three-dimensional reconstruction of the edge points of the wheel and the track, the traverse information is obtained in a non-contact method, which ensures the safety and improves the measurement accuracy.

3. The assistance of structured light greatly improves the efficiency of feature point matching and reduces the time complexity of the algorithm.

Description of drawings

FIG. 1 is a schematic diagram of the logic flow of the present invention.

Detailed ways

The present invention will be further described below in conjunction with specific embodiments.

As shown in Figure 1, the method for measuring the lateral movement of a train wheel provided by this embodiment includes the following steps:

平移因子其中，旋转因子

表示令左摄像头与右摄像头具有相同姿态的旋转效果，具体可以拆分为沿x、y、z轴分别旋转θ_x、θ_y、θ_z这3个步骤。即旋转因子中的参数r₁～r₉具体由3个旋转矩阵的乘积决定，Step 1: Use Zhang Zhengyou's calibration method to calibrate the binocular camera indoors to obtain the internal reference of the binocular camera and the relative position relationship matrix between the two cameras. Print a checkerboard calibration board with a known size, take 10 to 20 calibration board images with different positions, angles and attitudes, and extract sub-pixel corner coordinates for each calibration image for camera calibration. Obtain camera internal parameters: S _X , S _y , u ₀ , v ₀ , f, where S _X and S _y are the physical dimensions of a single pixel of the camera chip, u ₀ , v ₀ are the pixel coordinates of the optical center of the optical axis, and f is The focal length of the camera; the relative position relationship of the binocular camera is stereo calibrated, and the external parameters R and T are obtained, among which the rotation factor

translation factor where the twiddle factor

Represents the rotation effect of making the left camera and the right camera have the same posture, which can be divided into three steps of rotating θ _x , θ _y , and θ _z along the x, y, and z axes respectively. That is, the parameters r ₁ to r ₉ in the rotation factor are specifically determined by the product of the three rotation matrices,

The t _x , _ty , and t _z in the translation factor are the translation distances of the origin of the right camera relative to the left camera. Assuming that the coordinate system of the left camera is _XL and the coordinate system of the right camera is X _R , the relative formula of the two cameras can be obtained as X _R =RX _L +T. The stereo correction preprocessing is performed on the left and right images.

Step 2: Arrange a binocular camera and a structured light source on the train bogie, the shooting range includes the wheel and track to be tested, adjust the structured light source and project the linear structured light on the wheel and track to be tested.

Step 3: Perform probabilistic Hough line detection on the image, extract the structured light lines in the left view and the right view, and match the structured light segments in the left and right images one by one according to the angle of the line.

Step 4: Extract the image coordinates (X _l , Y _l ) and (X _r , Y _r ) of the same end point of the same line on the left and right cameras, and use the binocular stereo vision algorithm to calibrate the left and right camera internal parameters and external parameters R, T Calculate the three-dimensional space coordinates P(x _C , y _C , z _C ) of the feature points.

Step 5: Use 3 non-collinear points to determine a plane, the wheel edge points P ₁ , P ₂ , P ₃ have equal distances from the center of the circle and are all R, set the general constants A _i , B _i , C _i , D _i , where (A _i ,B _i ,C _i ) is the normal vector of the wheel plane, and the solution center O(x _i ,y _i ,z _i ) space coordinates:

A _i x+B _i y+C _i z+D _i =0

Assuming general constants m, n, and p, the equation of the space straight line l where the orbit is located is calculated by using the two characteristic points Q ₁ and Q ₂ on the same side of the orbit: Where (x ₀ , y ₀ , z ₀ ) is a point on the straight line l, and (m, n, p) is the direction vector of the straight line l. Assuming general constants A, B, C, D, using two feature points Q ₁ , Q ₂ on the same side of the track and one feature point Q ₃ on the other side, calculate the general equation of the plane α where the wheel is located: Ax+By +Cz+D=0, where (A, B, C) is the normal vector of plane α;

Step 6: Calculate the projection V(x, y, z) of the center O(x _i , y _i , _zi ) on the orbital plane α and the distance d between the projection point V and the orbital edge in turn. The vertical line of the orbital plane through the center O(x _i , y _i , z _i ) intersects the projection point V(x, y, z), and the parametric equation of the vertical line is obtained:

The parametric equation of the vertical line is brought into the orbit plane equation α, and the parameter t corresponding to the projection point V(x, y, z) in the parametric equation is obtained:

Then the parameter t is brought into the parametric equation of the straight line to obtain the specific spatial coordinates V(x, y, z) of the projection point. Using the two points Q ₁ and Q ₂ inside the track edge, find the distance from the projected point V to the track edge:

Step 7: Set the normal traverse range [d _min , d _max ], if the distance between the projection point of the circle center on the track plane and the inner edge of the track d _min <d < d _max , the train runs normally at this time; if the circle center is on the track If the distance between the projected point of the plane and the edge of the track exceeds the normal traverse range, the train runs abnormally at this time.

The above-mentioned embodiments are only preferred embodiments of the present invention, and are not intended to limit the scope of implementation of the present invention. Therefore, any changes made according to the shape and principle of the present invention should be included within the protection scope of the present invention.

Claims (4)

1. a measuring method of the lateral movement of a train wheel, is characterized in that, comprises the following steps: 1) A binocular camera and a structured light source are arranged on the train bogie. The shooting range includes the wheel and track to be tested. Adjust the structured light source and project the linear structured light on the wheel and track to be tested; System calibration, obtain the internal parameters of the binocular camera and perform stereo calibration to obtain the relative position relationship matrix between the binocular cameras, and use the internal and external parameters to perform stereo correction preprocessing on the left and right images; 2) Perform probabilistic Hough line detection on the image to extract the structured light lines in the left and right views; take the straight line angle as the matching feature, match the structured light segments in the left and right images one by one, and then match the same straight line segment The two end points of the two end points are matched; the binocular stereo vision algorithm is used to calculate the three-dimensional space coordinates of the feature points through the binocular internal parameters and external parameters obtained by calibration; 3) According to the spatial coordinates of the corresponding feature points, three-dimensional reconstruction of the wheel and the track is carried out, and the three-dimensional space equation of the wheel and the track is established; 4) Calculate the distance between the projection point of the wheel center on the track plane and the edge of the track as the lateral movement of the train wheel, and set the normal lateral movement range. If there is a distance between the projection point of the wheel center on the track plane and the edge of the track If the distance between the projection point of the wheel center on the track plane and the edge of the track exceeds the normal traverse range, it is determined that the train is running abnormally. 2. the measuring method of a kind of train wheel traverse according to claim 1, is characterized in that: in step 1), adopt binocular camera of parallel placement and a plurality of structured light sources, be placed on the train bogie, Adjust the angle of the camera and the structured light source, so that the binocular camera can capture images of the structured light irradiating on the wheel and the track at different angles. 3. the measuring method of a kind of train wheel lateral movement according to claim 1, is characterized in that: in step 1), adopt Zhang Zhengyou calibration method to demarcate binocular camera, obtain and comprise S _X , S _Y , _u . The internal parameters of , v ₀ , f, where S _X and S _Y are the physical dimensions of a single pixel of the camera chip, u ₀ and v ₀ are the pixel coordinates of the optical center of the optical axis, and f is the focal length of the camera; the relative position of the binocular camera The relationship is stereo calibrated, and the external parameters R and T are obtained, among which the twiddle factor

translation factor

twiddle factor Represents the rotation effect of making the left camera and the right camera have the same attitude, which can be divided into three steps of rotating θ _x , θ _y , and θ _z along the x, y, and z axes respectively, that is, the parameter r ₁ ~ r ₉ is specifically determined by the product of 3 rotation matrices:

t _x , _ty , and t _z in the translation factor are the translation distances of the origin of the right camera relative to the left camera. Let the coordinate system of the left camera be X _L and the coordinate system of the right camera be X _R , and the relative formula of the two cameras can be obtained as X _{R =} RXL+T. 4. the measuring method of a kind of train wheel lateral movement according to claim 1, it is characterized in that: in step 2), carry out probability Hough straight line detection to the image, detect the straight line that structured light is mapped on the image, select the structure The endpoints of the light on the edge of the object are used as feature points and matched. After calculation, three non-collinear feature points P ₁ (x ₁ , y ₁ , z ₁ ) and P ₂ (x ₂ , y ₂ , z ₂ are selected on the wheel. ), P ₃ (x ₂ , y ₂ , z ₂ ) and two feature points Q ₁ , Q ₂ on one side of the track and one feature point Q ₃ on the other side; In step 3), the three non-collinear points are used to determine a plane, the wheel edge points P ₁ , P ₂ , P ₃ have equal distances from the center of the circle and are all R, and set the constants A _i , B _i , C _i , D _i , where (A _i ,B _i ,C _i ) is the normal vector of the wheel plane, and the solution center O(x _i ,y _i ,z _i ) space coordinates: A _i x+B _i y+C _i z+D _i =0

Set constants m, n and p, and use the two characteristic points Q ₁ and Q ₂ on the same side of the orbit to calculate the equation of the space straight line l where the orbit is located:

Where (x ₀ , y ₀ , z ₀ ) is a point on the straight line l, (m, n, p) is the direction vector of the straight line l; set the constants A, B, C, D, use the 2 on the same side of the track The general equation of the plane α where the wheel is located is calculated: Ax+By+Cz+D _{= 0} , where ( _A , B, C) is the plane the normal vector of α; In step 4), calculate the projection _V (x,y, _z ) of the center of the circle on the orbital plane and the distance d between _V and the edge of the orbital; Intersecting at the projection point V(x, y, z), the parametric equation of the vertical line is obtained:

The parametric equation of the vertical line is brought into the orbit plane equation α, and the parameter t corresponding to the projection point V(x, y, z) in the parametric equation is obtained:

Then bring the parameter t into the parametric equation of the straight line to obtain the specific spatial coordinates V(x, y, z) of the projection point; use the two points Q ₁ and Q ₂ on the same side of the track edge to find the projection point V to the track edge the distance: Set the normal traverse range [d _min , d _max ]. If the distance between the projection point of the circle center on the track plane and the track edge is d _min <d < d _max , the train runs normally; if the circle center is at the projection point of the track plane If the distance from the edge of the track exceeds the normal traverse range, the train runs abnormally at this time.

Images