CN-121978893-A - Control method for camera gesture of high-speed rail contact net detection vehicle

CN121978893ACN 121978893 ACN121978893 ACN 121978893ACN-121978893-A

Abstract

The invention provides a control method of a camera attitude of a high-speed rail contact net detection vehicle, which relates to the technical field of high-speed rail contact net detection and comprises the steps of obtaining real-time speed, relative rolling angle and absolute rolling angle of a vehicle body of the high-speed rail contact net detection vehicle, wherein the relative rolling angle of the vehicle body is the inclination angle of the high-speed rail contact net detection vehicle relative to a high-speed rail, the absolute rolling angle of the vehicle body is the inclination angle of the high-speed rail contact net detection vehicle relative to an absolute horizontal plane, calculating a camera attitude error according to the relative rolling angle of the vehicle body and the absolute rolling angle of the vehicle body, calculating a camera attitude adjustment control parameter based on a vehicle body action PID control model, and adjusting the camera attitude of the high-speed rail contact net detection vehicle according to the camera attitude adjustment control parameter. The camera of the high-speed rail contact net detection vehicle can always keep absolute level, and the measurement accuracy of key geometric parameters such as the contact net height guiding and pulling-out values is ensured.

Inventors

QU AIHUA
KANG JUNWEI
Xu futian
LEI CONGYI
SHI JIXING

Assignees

河北工业大学

Dates

Publication Date: 20260505
Application Date: 20260210

Claims (10)

1. The control method for the camera gesture of the high-speed rail contact net detection vehicle is characterized by comprising the following steps of: S1, acquiring real-time speed, relative rolling angle and absolute rolling angle of a vehicle body of a high-speed railway contact net detection vehicle, wherein the relative rolling angle of the vehicle body is the inclination angle of the high-speed railway contact net detection vehicle relative to a high-speed railway track; s2, calculating a camera attitude error according to the relative roll angle of the vehicle body and the absolute roll angle of the vehicle body; And S3, calculating camera attitude adjustment control parameters based on a car body action PID control model according to the camera attitude error, the car body absolute roll angle and the real-time speed, and adjusting the attitude of the high-speed rail contact net detection car camera according to the camera attitude adjustment control parameters.
2. The method for controlling the camera attitude of a high-speed railway contact net detection vehicle according to claim 1, wherein obtaining the relative roll angle of the vehicle body of the high-speed railway contact net detection vehicle comprises: the method comprises the steps of obtaining a first pair of rail distances and a second pair of rail distances of a high-speed rail contact net detection vehicle, wherein the first pair of rail distances are distances between a first position on the bottom surface of the high-speed rail contact net detection vehicle and a high-speed rail track, and the second pair of rail distances are distances between a second position on the bottom surface of the high-speed rail contact net detection vehicle and the high-speed rail track; and calculating the relative roll angle of the vehicle body according to the first pair of rail distances, the second pair of rail distances and the preset distance.
3. The method for controlling the camera pose of a high-speed railway contact net detection vehicle according to claim 1, wherein obtaining the absolute roll angle of the body of the high-speed railway contact net detection vehicle comprises: Collecting first acceleration and second acceleration of a high-speed rail contact net detection vehicle through an accelerometer, wherein the first acceleration is acceleration of the high-speed rail contact net detection vehicle along an absolute horizontal direction perpendicular to a running direction of the high-speed rail contact net detection vehicle, and the second acceleration is acceleration of the high-speed rail contact net detection vehicle perpendicular to the absolute horizontal direction; And calculating the absolute roll angle of the vehicle body according to the first acceleration and the second acceleration.
4. The method for controlling camera attitude of a high-speed railway catenary detection vehicle according to claim 3, wherein calculating a camera attitude error from the relative roll angle of the vehicle body and the absolute roll angle of the vehicle body comprises: If the relative roll angle of the vehicle body is smaller than or equal to a first preset angle threshold value, calculating a camera inclination angle difference value according to the absolute roll angle of the vehicle body, and taking the camera inclination angle difference value as the camera attitude error; if the relative roll angle of the vehicle body is larger than a first preset angle threshold, calculating a predicted value of the absolute roll angle of the vehicle body based on a gyroscope state prediction model, calculating a zero drift correction value of the absolute roll angle of the vehicle body based on a Kalman filtering fusion algorithm according to the predicted value, and taking the zero drift correction value as the camera attitude error.
5. The method for controlling the camera pose of a high-speed railway catenary detection vehicle according to claim 4, wherein calculating the predicted value of the absolute roll angle of the vehicle body based on a gyroscope state prediction model comprises: Acquiring the angular speed of a high-speed rail contact net detection vehicle through a gyroscope, and acquiring an absolute roll angle of a vehicle body at the last moment and a zero offset of the gyroscope at the last moment; and inputting the absolute roll angle of the vehicle body at the previous moment, the zero offset of the gyroscope at the previous moment and the angular speed of the high-speed rail contact net detection vehicle into a gyroscope state prediction model to obtain a predicted value of the absolute roll angle of the vehicle body.
6. The control method for the camera pose of the high-speed railway overhead line system detection vehicle according to claim 4, wherein obtaining the zero drift correction value of the absolute roll angle of the vehicle body based on a kalman filter fusion algorithm according to the predicted value comprises: calculating an initial observation value of the absolute roll angle of the vehicle body based on a Kalman filtering observation standard equation according to the prediction value and the relative roll angle of the vehicle body; according to the first acceleration, calculating an observation noise variance, and dynamically adjusting a weight parameter of the Kalman filtering observation standard equation according to the observation noise variance to obtain a Kalman filtering observation target equation; and calculating a zero drift correction value of the absolute roll angle of the vehicle body based on the Kalman filtering observation target equation according to the predicted value and the relative roll angle of the vehicle body.
7. The method for controlling the camera pose of a high-speed rail overhead line system detection vehicle according to claim 4, wherein calculating the camera pose adjustment control parameter based on the vehicle body motion PID control model comprises: Generating a vehicle body state vector according to the attitude error, the absolute roll angle of the vehicle body and the real-time speed; Inputting the vehicle body state vector into a preset action network sub-model to obtain a camera action vector; Calculating camera motion control parameters according to the camera motion vector; And inputting the camera motion control parameters and the attitude errors into an incremental PID calculation sub-model to obtain the camera attitude adjustment control parameters.
8. The method for controlling the camera pose of the high-speed rail contact network detection vehicle according to claim 7, wherein the camera motion control parameters comprise a camera response sensitivity coefficient, a camera steady-state integral coefficient and a camera anti-shake differential coefficient.
9. The method for controlling the camera pose of a high-speed rail contact network inspection vehicle according to claim 2, wherein obtaining a first pair of rail distances and a second pair of rail distances of the high-speed rail contact network inspection vehicle comprises: obtaining the distance between high-speed rail sleeper; calculating a noise fundamental frequency according to the Gao Tiezhen wood spacing and the real-time speed; the center frequency of the dynamic notch filter is the noise fundamental frequency, and the bandwidth of the dynamic notch filter is determined according to the fluctuation condition of the real-time speed and the construction error of the Gao Tiezhen wood spacing; collecting a measured value corresponding to the first pair of track distances and a measured value corresponding to the second pair of track distances; and inputting the measured values corresponding to the first pair of track distances and the measured values corresponding to the second pair of track distances into a dynamic notch filter to obtain the first pair of track distances and the second pair of track distances.
10. The method for controlling the camera pose of a high-speed rail contact network inspection vehicle according to claim 7, further comprising: Acquiring a camera image blurring punishment parameter, a camera control energy consumption punishment parameter and a camera tracking precision punishment parameter; Calculating a camera action rewarding parameter according to the camera image blurring punishment parameter, the camera control energy consumption punishment parameter and the camera tracking precision punishment parameter; and dynamically adjusting the network weight of the preset action network sub-model according to the camera action rewarding parameter.

Description

Control method for camera gesture of high-speed rail contact net detection vehicle Technical Field The invention relates to the technical field of high-speed rail contact net detection, in particular to a control method for camera gestures of a high-speed rail contact net detection vehicle. Background In the running process of the high-speed rail contact net detection vehicle, a camera installed on the roof or the bottom of the vehicle is required to shoot key geometric parameters (such as a guide height and a pull-out value) of the contact net. In order to guarantee the geometric validity of the measurement data, it is required that the camera viewing axis must always remain parallel to the absolute horizontal plane (i.e. perpendicular to the direction of gravity). In the prior art, a laser radar-based triangulation ranging method is generally adopted to monitor whether a vehicle body is parallel to a rail surface so as to adjust the angle of a camera, or a conventional PID-controlled cradle head is adopted to forcedly enable the camera to be parallel to an absolute horizontal plane. However, when passing through a curve, the track itself is "over-height" (i.e. the outer rail is higher than the inner rail) when the body is tilted with the track. The traditional laser radar scheme can only detect the relative distance between the vehicle body and the rail surface, and the relative distance difference is 0 at this moment, and the system can be misjudged as 'horizontal', so that the camera inclines along with the vehicle body, the photographed picture rotates, and the measurement accuracy of the subsequent geometric parameters is seriously affected. The fixed parameters of the conventional PID control holder cannot be adapted to nonlinear vibration characteristics of different working conditions such as high speed of a train, curves and the like, so that response speed and control stability are difficult to be considered, and further measurement accuracy of subsequent geometric parameters is affected. Therefore, a control method for the camera gesture of the high-speed rail contact net detection vehicle is needed to improve the capability of keeping absolute level of the camera under different working conditions of the high-speed rail contact net detection vehicle operation, and ensure the measurement precision of key geometric parameters such as the contact net height guiding and the pull-out value. Disclosure of Invention The invention aims to solve the problems and provides a control method for the camera gesture of a high-speed rail contact net detection vehicle. The invention provides a control method for the camera gesture of a high-speed rail contact net detection vehicle, which comprises the following steps: S1, acquiring real-time speed, relative rolling angle and absolute rolling angle of a vehicle body of a high-speed railway contact net detection vehicle, wherein the relative rolling angle of the vehicle body is the inclination angle of the high-speed railway contact net detection vehicle relative to a high-speed railway track; s2, calculating a camera attitude error according to the relative roll angle of the vehicle body and the absolute roll angle of the vehicle body; And S3, calculating camera attitude adjustment control parameters based on a car body action PID control model according to the camera attitude error, the car body absolute roll angle and the real-time speed, and adjusting the attitude of the high-speed rail contact net detection car camera according to the camera attitude adjustment control parameters. According to the technical scheme provided by some embodiments of the invention, the method for obtaining the relative roll angle of the high-speed rail contact net detection vehicle body comprises the following steps: the method comprises the steps of obtaining a first pair of rail distances and a second pair of rail distances of a high-speed rail contact net detection vehicle, wherein the first pair of rail distances are distances between a first position on the bottom surface of the high-speed rail contact net detection vehicle and a high-speed rail track, and the second pair of rail distances are distances between a second position on the bottom surface of the high-speed rail contact net detection vehicle and the high-speed rail track; and calculating the relative roll angle of the vehicle body according to the first pair of rail distances, the second pair of rail distances and the preset distance. According to the technical scheme provided by some embodiments of the invention, the method for obtaining the absolute roll angle of the body of the high-speed rail contact net detection vehicle comprises the following steps: The method comprises the steps that a first acceleration and a second acceleration of a high-speed rail contact net detection vehicle are collected through an accelerometer, wherein the first acceleration is the acceleration of the high-speed rail contact net de