A wire ranging method for a bucket arm vehicle based on active search

Abstract

The application provides a wire ranging method for a bucket arm vehicle based on active search, which is suitable for accurate ranging of a wire in a complex environment. The method integrates a high-precision holder device on the basis of traditional binocular vision, constructs an active binocular vision detection system, takes the insulating terminals on the left and right sides of the wire (connection points of the wire and a wire pole) as detection targets, establishes a dynamic space geometric relationship model by acquiring three-dimensional coordinates of the insulating terminals and holder pose information in real time, and realizes high-precision real-time calculation of the spatial position of the wire. The method effectively solves the technical problem of inaccurate power line sensing of the bucket arm vehicle in a complex operation environment, and provides reliable technical support for safe operation of special vehicles.

Description

Technical Field

[0001] This invention relates to an active search-based method for power line ranging in boom trucks, belonging to the field of special vehicle safety operation technology. Specifically, it relates to a three-dimensional sensing system and method for power lines in boom truck operation scenarios. Background Technology

[0002] With the acceleration of urbanization, power lines are ubiquitous in urban infrastructure. In such environments, the need for construction, maintenance, and inspection using boom lifts and other intelligent equipment is increasing. Failure to effectively identify and avoid power lines during equipment operation can lead to collisions, damage, or even safety accidents. Therefore, accurately detecting and avoiding power lines is crucial for maintenance work in urban environments.

[0003] Vision-based power line sensing technologies include monocular vision and binocular vision methods. While monocular vision systems offer advantages such as simple information acquisition and fast computation speed, they often struggle to accurately identify power lines in complex urban environments, especially when power lines highly overlap with other objects or when visual interference exists, leading to misjudgments. Binocular vision systems offer stronger spatial perception capabilities compared to monocular vision and can effectively improve the accuracy of power line identification. However, due to the presence of sparse and feature-sparse power lines in the working environment, traditional binocular vision systems struggle with feature point matching when identifying and measuring distances from power lines, thus failing to accurately interpret distance information and meet the perception accuracy requirements of special working environments.

[0004] Therefore, this invention proposes a method for measuring the distance of power lines on a bucket truck based on active search. By adding a gimbal to optimize detection performance, it improves the safety, efficiency, and intelligence of automated equipment operating near power lines, solves the problem of inaccurate detection of power lines in special environments, enhances environmental adaptability, reduces potential operational risks, and promotes the application of intelligent equipment in urban infrastructure management. Summary of the Invention

[0005] This invention addresses the limitation of bucket trucks in accurately sensing power lines under special working conditions by providing an active search-based method for measuring the distance of power lines on bucket trucks. The method involves adding a gimbal to a binocular vision system to detect the insulated terminals of the power lines on both sides. By combining the angle of rotation of the active camera towards the insulated terminals on both sides, geometric relationships are used to achieve accurate distance measurement of the power lines.

[0006] This invention is a method for power line ranging on a bucket truck based on active search, comprising the following steps: S1: Add an active gimbal to the traditional binocular camera to calibrate the parameters of the binocular camera; S2: Detect wires using a binocular camera and number and match the detected wire segments; S3: The binocular camera tracks the detected lines to locate the insulating terminal and obtain the distance between the insulating terminal and the camera; stereo matching is performed on the images acquired on the left and right sides to generate a parallax image; the depth of the target point is calculated based on the parallax, and the three-dimensional spatial coordinates of the insulating terminals A and B are solved. S4: Calculate the angle of the left insulating terminal A in the triangle based on the gimbal rotation angle, the distance between the two insulating terminals A and B, and the distance between the right insulating terminal B and the camera; then calculate the distance between the camera and the power line based on the angle value and the distance between the left insulating terminal A and the camera.

[0007] Furthermore, S1 includes the following steps: S11: Install a binocular camera with an active gimbal on the bucket truck; S12: Using the Zhang Zhengyou calibration method, the system completes precise calibration and parameter calibration by taking pictures of the standard checkerboard pattern from multiple angles, obtaining the intrinsic parameter matrix of the left and right cameras, the extrinsic parameter transformation matrix between the two cameras, and the radial and tangential distortion coefficients of the lens.

[0008] Furthermore, S2 includes the following steps: S21: Make the optical axes of the left and right cameras parallel, and the left and right cameras complete the exposure and image acquisition at the same time, ensuring that the two images correspond to the scene state at the same time. S22: Extract the pixel contours of thin line targets such as wires from the images captured by the left and right cameras respectively using image algorithms; S23: After the left and right cameras detect the lines, a vertical coordinate sorting system is established with the left boundary of the image as the reference.

[0009] Furthermore, S22 includes the following steps: S221: Preprocessing: The acquired image is denoising through Gaussian filtering, contrast is enhanced through histogram equalization, and the edge contour of the target is accurately extracted from the image using the Canny algorithm. Image noise is suppressed, strong edges are filtered, and weak edges are connected. S221: Line Extraction: The probabilistic Hough transform is used to detect wire lines, and the pixel regions of the wires are separated from the background to obtain the set of pixel coordinates of each wire in the image, such as the pixel sequence of the start point, end point, and center line. S221: Target Filtering: Based on the aspect ratio, length, position, and other characteristics of the lines, remove interfering lines in the background, such as tree branches and railings, and retain only the target wires.

[0010] Furthermore, S23 includes the following steps: S231: For all detected line segments, calculate their y-coordinate at the horizontal line (x=L / 2) at the center of the image; S232: Compare the y coordinates, sort them in ascending order according to the y coordinate values, and number them; S233: For the i-th line segment in the left camera image, find matching line segments in the right camera image that meet the following conditions, and establish a database of one-to-one correspondence between lines in the left and right cameras; The conditions include: The numbers are the same; The difference in line inclination angle is less than 3 degrees; The y-coordinate difference at the center horizontal line is less than ±10% of the theoretical parallax corresponding to the baseline distance.

[0011] Furthermore, S3 includes the following steps: S31: The left camera tracks the detected line on the left side, calculating the average gradient magnitude within the tracking window in real time. When the gradient magnitude of three consecutive tracking points exceeds a threshold... ( When μ is the mean of the wire's background gradient and σ is the standard deviation, it indicates that the tracking is in an ideal state and is determined to be an insulated terminal A. S32: Then the right camera tracks the detected line to the left, performing the same operation until the right camera also detects the insulated terminal A; S33: At this point, the three-dimensional coordinates of the insulated terminal A are calculated using parallax to obtain its distance from the camera. ; S34: Similarly, the left and right cameras track and locate the insulated terminal B along the right side of the wire, completing its three-dimensional coordinate measurement and obtaining the distance. .

[0012] Furthermore, S4 includes the following steps: S41: Based on the measured distances between insulated terminals A and B and the camera. , And the angle between insulating terminal A, camera, and insulating terminal B. This fixes the triangle formed by the left terminal, the camera, and the right terminal. ; S42: In the triangle formed by the insulating terminals A and B and the camera's location point C, according to the law of sines, we can obtain... , By combining the two formulas, the distance d from the camera to the power line can be obtained.

[0013] Through the above design scheme, the present invention can bring the following beneficial effects.

[0014] 1. By actively searching with the left and right cameras, it can be ensured that the wire insulation terminals are always within the effective measurement field of view of the binocular system. This technology fundamentally solves the measurement failure problem caused by the insulation terminals being outside the field of view in traditional binocular vision.

[0015] 2. This invention uses insulated terminals as feature detection targets, and directly locates feature points of insulated terminals on both sides of the wire through active binocular collaborative search, avoiding the problem of wire feature matching in complex backgrounds. This method significantly improves the reliability and environmental adaptability of the system, and is particularly suitable for wire ranging scenarios in complex backgrounds.

[0016] 3. This invention establishes a geometric model for wire ranging by obtaining three-dimensional coordinate measurements, and can output the distance from any position of the wire to the camera. Attached Figure Description

[0017] The present invention will be further described below with reference to the accompanying drawings: Figure 1 The flowchart of the active search-based electric wire ranging method for bucket boom trucks is as follows: Figure 2 This is the active search model for power line ranging on a bucket truck according to the present invention. Detailed Implementation

[0018] The present invention will now be described in detail with reference to the accompanying drawings.

[0019] The flowchart of the electric wire ranging method for bucket boom trucks based on active search is as follows: Figure 1 As shown, the specific steps include the following: S1: Add an active gimbal to the traditional binocular camera to calibrate the parameters of the binocular camera; S2: Detect wires using a binocular camera and number and match the detected wire segments; S3: The binocular camera tracks the detected lines to locate the insulating terminal and obtain the distance between the insulating terminal and the camera; stereo matching is performed on the images acquired on the left and right sides to generate a parallax image; the depth of the target point is calculated based on the parallax, and the three-dimensional spatial coordinates of the insulating terminals A and B are solved. S4: Calculate the angle of the left insulating terminal A in the triangle based on the gimbal rotation angle, the distance between the two insulating terminals A and B, and the distance between the right insulating terminal B and the camera; then calculate the distance between the camera and the power line based on the angle value and the distance between the left insulating terminal A and the camera.

[0020] S1 includes the following steps: S11: A binocular camera with an active gimbal is mounted on a boom truck. The gimbal is driven by a motor and can actively adjust its attitude, dynamically adjusting the camera's pose. This solves the problem of traditional binocular cameras having "fixed viewing angle and being susceptible to shaking interference," thereby improving the stability, flexibility, and accuracy of 3D perception. The system employs an MS9015 V1 three-phase permanent magnet synchronous motor as the gimbal drive, boasting excellent performance specifications of 0.8 N·m rated torque and 130 rpm rated speed. A control system is built around an STM32F103VET6 microcontroller, utilizing an optimized space vector pulse width modulation (SVPWM) algorithm to achieve efficient motor drive. A real-time data exchange channel between the controller and motor is established using a CAN bus communication protocol, with a communication baud rate set to 1 Mbps to ensure rapid response and precise execution of rotation control commands. The system integrates a 17-bit high-precision encoder, forming a closed-loop control architecture that achieves an angular positioning accuracy of ±0.05°.

[0021] S12: Using the Zhang Zhengyou calibration method, the system completes precise calibration and parameter calibration by taking pictures of the standard checkerboard pattern from multiple angles, obtaining the intrinsic parameter matrix of the left and right cameras, the extrinsic parameter transformation matrix between the two cameras, and the radial and tangential distortion coefficients of the lens.

[0022] Furthermore, S2 includes the following steps: S21: Make the optical axes of the left and right cameras parallel, and the left and right cameras complete the exposure and image acquisition at the same time, ensuring that the two images correspond to the scene state at the same time. S22: Extract the pixel contours of thin line targets such as wires from the images captured by the left and right cameras respectively using image algorithms; S23: After the left and right cameras detect the lines, a vertical coordinate sorting system is established with the left boundary of the image as the reference.

[0023] Furthermore, S22 includes the following steps: S221: Preprocessing: The acquired image is denoising through Gaussian filtering, contrast is enhanced through histogram equalization, and the edge contour of the target is accurately extracted from the image using the Canny algorithm. Image noise is suppressed, strong edges are filtered, and weak edges are connected. S221: Line Extraction: The probabilistic Hough transform is used to detect wire lines, and the pixel regions of the wires are separated from the background to obtain the set of pixel coordinates of each wire in the image, such as the pixel sequence of the start point, end point, and center line. S221: Target Filtering: Based on the aspect ratio, length, position, and other characteristics of the lines, remove interfering lines in the background, such as tree branches and railings, and retain only the target wires.

[0024] Among these methods, the probability Hough transform is used to remove lines that do not meet the requirements. Set the length threshold to ≥50 pixels; this means that the wire occupies at least 50 pixels in length from beginning to end in the image. The aspect ratio threshold is ≥20:1; this means that the ratio of the length to the width of the wire must be greater than or equal to 20, and the core is to screen out slender lines. Location thresholds (midpoint): x∈[100,1180], y∈[50,670]; Angle threshold: ±5°; This indicates that the wire runs completely horizontally, and the core function is to filter out horizontal lines.

[0025] Furthermore, S23 includes the following steps: S231: For all detected line segments, calculate their y-coordinate at the horizontal line (x=L / 2) at the center of the image; S232: Compare the y coordinates, sort them in ascending order according to the y coordinate values, and number them; S233: For the i-th line segment in the left camera image, find matching line segments in the right camera image that meet the following conditions, and establish a database of one-to-one correspondence between lines in the left and right cameras; The conditions include: The numbers are the same; The difference in line inclination angle is less than 3 degrees; The y-coordinate difference at the center horizontal line is less than ±10% of the theoretical parallax corresponding to the baseline distance.

[0026] Furthermore, S3 includes the following steps: S31: The left camera tracks the detected line on the left side, calculating the average gradient magnitude within the tracking window in real time. When the gradient magnitude of three consecutive tracking points exceeds a threshold... ( When μ is the mean of the wire's background gradient and σ is the standard deviation, it indicates that the tracking is in an ideal state and is determined to be an insulated terminal A. S32: Then the right camera tracks the detected line to the left, performing the same operation until the right camera also detects the insulated terminal A; Stereo matching is performed on the acquired image of insulating terminal A to generate a parallax image, and then the parallax is used to... Calculate the depth of the target point Solve the three-dimensional spatial coordinates of insulating terminal A. , , .

[0027] in, Here, B represents the camera's optical center coordinates, which can be obtained through calibration; B represents the baseline distance of the stereo cameras. The focal length is used; similarly, the left and right cameras track and locate the insulating terminal B along the right side of the wire, and calculate the depth and three-dimensional coordinates of the insulating terminal B. .

[0028] S33: At this point, the three-dimensional coordinates of the insulated terminal A are calculated using parallax to obtain its distance from the camera. ; S34: Similarly, the left and right cameras track and locate the insulated terminal B along the right side of the wire, completing its three-dimensional coordinate measurement and obtaining the distance. .

[0029] Furthermore, S4 includes the following steps: S41: Based on the measured distances between insulated terminals A and B and the camera. , And the angle between insulating terminal A, camera, and insulating terminal B. This fixes the triangle formed by the left terminal, the camera, and the right terminal. ; The calculation process for the distance between insulating terminals A and B is as follows: the distance between insulating terminals A and B and the camera is obtained by binocular distance measurement. , , in,

[0030] in,

[0031]

[0032] in,

[0033] These represent the coordinate differences between the two ends of the wire in the X, Y, and Z axes of the camera coordinate system.

[0034] S42: In the triangle formed by the insulating terminals A and B and the camera's location point C, according to the law of sines, we can obtain... , By combining the two formulas, the distance d from the camera to the power line can be obtained.

[0035] . The left and right cameras are denoted as l and r, respectively.

Claims

1. A method for measuring the distance of overhead power lines on a bucket truck based on active search, characterized in that, Includes the following steps: S1: Add an active gimbal to the traditional binocular camera to calibrate the parameters of the binocular camera; S2: Detect wires using a binocular camera and number and match the detected wire segments; S3: The binocular camera tracks the detected lines to locate the insulating terminal and obtain the distance between the insulating terminal and the camera; stereo matching is performed on the images acquired on the left and right sides to generate a parallax image; the depth of the target point is calculated based on the parallax, and the three-dimensional spatial coordinates of the insulating terminals A and B are solved. S4: Calculate the angle of the left insulating terminal A in the triangle based on the gimbal rotation angle, the distance between the two insulating terminals A and B, and the distance between the right insulating terminal B and the camera; then calculate the distance between the camera and the power line based on the angle value and the distance between the left insulating terminal A and the camera.

2. The method for power line ranging on a bucket boom truck based on active search according to claim 1, characterized in that, S1 includes the following steps: S11: Install a binocular camera with an active gimbal on the bucket truck; S12: Using the Zhang Zhengyou calibration method, the system completes precise calibration and parameter calibration by taking pictures of the standard checkerboard pattern from multiple angles, obtaining the intrinsic parameter matrix of the left and right cameras, the extrinsic parameter transformation matrix between the two cameras, and the radial and tangential distortion coefficients of the lens.

3. The method for measuring the distance of overhead power lines for bucket trucks based on active search according to claim 1, characterized in that, S2 includes the following steps: S21: Make the optical axes of the left and right cameras parallel, and the left and right cameras complete the exposure and image acquisition at the same time, ensuring that the two images correspond to the scene state at the same time. S22: Extract the pixel contours of thin line targets such as wires from the images captured by the left and right cameras respectively using image algorithms; S23: After the left and right cameras detect the lines, a vertical coordinate sorting system is established with the left boundary of the image as the reference.

4. The method for measuring the distance of overhead power lines for bucket trucks based on active search according to claim 3, characterized in that, S22 includes the following steps: S221: Preprocessing: The acquired image is denoising through Gaussian filtering, contrast is enhanced through histogram equalization, and the edge contour of the target is accurately extracted from the image using the Canny algorithm. Image noise is suppressed, strong edges are filtered, and weak edges are connected. S221: Line Extraction: The probabilistic Hough transform is used to detect wire lines, and the pixel regions of the wires are separated from the background to obtain the set of pixel coordinates of each wire in the image, such as the pixel sequence of the start point, end point, and center line. S221: Target Filtering: Based on the aspect ratio, length, position, and other characteristics of the lines, remove interfering lines in the background, such as tree branches and railings, and retain only the target wires.

5. The method for measuring the distance of overhead power lines for bucket trucks based on active search according to claim 3, characterized in that, S23 includes the following steps: S231: For all detected line segments, calculate their y-coordinate at the horizontal line (x=L / 2) at the center of the image; S232: Compare the y coordinates, sort them in ascending order according to the y coordinate values, and number them; S233: For the i-th line segment in the left camera image, find matching line segments in the right camera image that meet the following conditions, and establish a database of one-to-one correspondence between lines in the left and right cameras; The conditions include: The numbers are the same; The difference in line inclination angle is less than 3 degrees; The y-coordinate difference at the center horizontal line is less than ±10% of the theoretical parallax corresponding to the baseline distance.

6. The method for power line ranging on a bucket boom truck based on active search according to claim 1, characterized in that, S3 includes the following steps: S31: The left camera tracks the detected line on the left side, calculating the average gradient magnitude within the tracking window in real time. When the gradient magnitude of three consecutive tracking points exceeds a threshold... ( When μ is the mean of the wire's background gradient and σ is the standard deviation, it indicates that the tracking is in an ideal state and is determined to be an insulated terminal A. S32: Then the right camera tracks the detected line to the left, performing the same operation until the right camera also detects the insulated terminal A; S33: At this point, the three-dimensional coordinates of the insulated terminal A are calculated using parallax to obtain its distance from the camera. ; S34: Similarly, the left and right cameras track and locate the insulated terminal B along the right side of the wire, completing its three-dimensional coordinate measurement and obtaining the distance. .

7. The method for measuring the distance of overhead power lines for bucket trucks based on active search according to claim 1, characterized in that, S4 includes the following steps: S41: Based on the measured distances between insulated terminals A and B and the camera. , And the angle between insulating terminal A, camera, and insulating terminal B. This fixes the triangle formed by the left terminal, the camera, and the right terminal. ; S42: In the triangle formed by the insulating terminals A and B and the camera's location point C, according to the law of sines, we can obtain... , By combining the two formulas, the distance d from the camera to the power line can be obtained.

Images