Circuit breaker operating characteristic measurement method based on binocular vision and improved optical flow method

Abstract

The invention provides a circuit breaker operation characteristic measurement method based on binocular vision and an improved optical flow method, and the method comprises the steps: obtaining video data of the switching-on and switching-off operation process of a circuit breaker through employing a dual-camera camera, and detecting a target point in a first frame; corresponding feature points are found in the image based on the double-camera-position camera; tracking the pixel coordinate change of the feature points in the image sequence by adopting an improved optical flow method; and calculating the real three-dimensional coordinates of the feature points in the image sequence, and establishing a three-dimensional travel curve.

Description

Technical Field

[0001] The invention belongs to the technical field of physical detection, and in particular relates to a circuit breaker action characteristic measurement method based on binocular vision and an improved optical flow method. Background Art

[0002] Circuit breakers are key protection devices in power systems, and their opening and closing performance directly affects the safety and reliability of the system. The opening and closing curve (i.e., the curve of the moving contact displacement changing over time) is an important indicator for evaluating the mechanical performance of circuit breakers, and can reflect information such as the opening and closing speed characteristics, acceleration characteristics, and buffering effect of circuit breakers. By analyzing the curve, the mechanical state of the circuit breaker can be evaluated and its potential faults can be diagnosed.

[0003] Traditional circuit breaker action measurement methods rely on mechanical sensors or manual measurement, which are easily affected by environmental factors, resulting in low measurement accuracy. In addition, manual measurement and mechanical sensor methods are complex and time-consuming. Mechanical sensors need to directly contact the circuit breaker, which may cause interference or damage to the equipment.

[0004] Chinese patent application document CN106526467B records a method for measuring the opening and closing speed characteristics of a high-voltage circuit breaker based on machine vision. While the high-voltage circuit breaker is in operation, a high-speed camera is used to photograph the color marks set on the surface of the moving parts it observes. The recorded video images are processed and recognized through a machine vision algorithm to determine the position coordinates of the object observed in each frame of the video relative to the reference point, obtain the motion trajectory of the observed object, and then calculate the opening and closing speeds of the moving contacts of the high-voltage circuit breaker.

[0005] The existing measurement methods can only obtain the motion characteristics of the circuit breaker opening and closing process in two-dimensional space, but the circuit breaker is actually a component working in three-dimensional space. One-dimensional and two-dimensional image detection cannot fully reflect the dynamic characteristics of the circuit breaker contacts in three-dimensional space, and the test results are not comprehensive. Summary of the invention

[0006] In order to solve the above technical problems, this application proposes a circuit breaker action characteristic measurement method based on binocular vision and improved optical flow method. The specific technical solution is as follows: A circuit breaker action characteristic measurement method based on binocular vision and improved optical flow method, the method is: Use dual-position cameras to obtain video data of the circuit breaker opening and closing process and detect the target point in the first frame; Find the corresponding feature points in the image based on the dual-position camera; An improved optical flow method is used to track the pixel coordinate changes of feature points in image sequences; Calculate the true three-dimensional coordinates of the feature points in the image sequence and establish a three-dimensional travel curve.

[0007] Furthermore, the method of using the improved optical flow method to track the pixel coordinate changes of feature points in the image sequence is as follows: Establish an improved optical flow model function; Simplify the image parameterization function in the optical flow model function to obtain the Jacobian matrix and Hessian matrix; Introduce adaptive step size to accelerate iteration and add attenuation coefficient Control access to historical information; Repeat the iteration until convergence, then determine that the point is the exact position of the tracking point in this frame, and start searching for the target position of the next frame with this point as the initial coordinate; Repeat the above steps until the last frame.

[0008] Furthermore, the improved optical flow model function is: Where L is the loss function, is the gray value of the i-th frame image, T is the initial frame The extracted sub-region containing the target features, is a function of image parameterization, x is a matrix including the horizontal and vertical coordinates of the image, and v is the pixel velocity of the target area in the image.

[0009] Furthermore, the function Simplified to , get the Jacobian matrix and the Hessian matrix .

[0010] Furthermore, the attenuation coefficient ; In the above formula, is the cumulative square acceleration; is the historical information preservation degree, the value is (0,1); Acceleration step length ; Update speed: ; In the above formula, a is the adaptive step size, is the total step length; It is a hyperparameter, which is given manually. First, it prevents the denominator in the calculation formula of the adaptive step size a from being zero, resulting in an undefined adaptive step size a. Second, it controls the convergence speed of the adaptive step size by updating the calculation formula of the speed V.

[0011] Further, iterate until convergence, .

[0012] Furthermore, the real coordinate value of the target point is defined as By solving the following formula:

[0013] In the above formula, and is the pixel coordinate of the target point P in the two cameras; The conversion formula between pixel coordinates and world coordinates in the two cameras is: ; ; In the above formula, and is the camera’s intrinsic and extrinsic matrix, and are the Z coordinates of point P in the two camera coordinate systems respectively.

[0014] The beneficial effects of the present invention are: (1) Using binocular vision measurement principle, a three-dimensional travel curve of the circuit breaker opening and closing is established, which can obtain the position change of the circuit breaker in the depth direction and provide richer data support for circuit breaker status monitoring; (2) The improved LK optical flow method is used for tracking to improve measurement efficiency and ensure the real-time measurement results. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 What is shown is a flow chart of a method for measuring the opening and closing action characteristics of a circuit breaker based on binocular vision and improved optical flow method; Figure 2 Shown is a schematic diagram of a binocular camera model. DETAILED DESCRIPTION

[0016] In the following description, certain specific details are set forth in order to provide a thorough understanding of various embodiments. However, it will be appreciated by those skilled in the art that the present invention may be practiced without these details. In other cases, well-known structures are not shown or described in detail to avoid unnecessarily obscuring the description of the embodiments. Unless the context otherwise requires, throughout the specification and the appended claims, the word "comprising" shall be interpreted in an open, inclusive sense, i.e., as "including but not limited to".

[0017] For the opening and closing action process of the circuit breaker, the action characteristics are measured in the following way: (1) Set image acquisition conditions; (2) Obtain video data of the circuit breaker opening and closing process, and detect corner points in the first frame of the video data as the basis for subsequent tracking and matching; (3) Under the epipolar constraint, use the stereo matching algorithm to find the corresponding feature points in the image; (4) Use the improved optical flow method to track the pixel coordinate changes of feature points in the image sequence; (5) Calculate the true three-dimensional coordinates of the feature points and establish a three-dimensional travel curve.

[0018] In the implementation method, images are acquired using a dual-camera binocular camera. The two binocular cameras should be placed in parallel and at the same height. The parameters of the binocular cameras are set to 720p pixels and a shooting frame rate ≥ 1000 frames / second to ensure that the action process of the circuit breaker can be clearly captured.

[0019] The binocular camera should be calibrated in advance using a calibration plate. High-contrast checkerboard marking stickers should be pasted on the motion link. Multiple images should be taken and the internal and external parameters of the camera should be calculated using the Zhang Zhengyou calibration method.

[0020] In step (2), the Shi-Tomasi corner point detection method is used to detect corner points in the first frame image. Corner points are a type of feature points. Feature points include corner points, spots, and other regions with features.

[0021] In step (3), the internal and external parameters of the binocular camera are first obtained through camera calibration, and the epipolar equation is calculated. The corresponding epipolar line is found in the right image according to the epipolar equation for the feature points detected in the images of the two binocular cameras, and the matching search is performed only on the epipolar line, thereby narrowing the search range; then a semi-global matching algorithm is used to determine the correspondence between the feature points in the left and right images by calculating the disparity map.

[0022] The basic premise of using the optical flow method is that the grayscale value of the tracked target remains consistent between different frames. This method uses the least squares method for template matching to determine an optimal speed so that the target matches the initially extracted template.

[0023] In step (4), by simplifying the calculation of the Hessian matrix, the amount of calculation required for each iteration is effectively reduced to maintain high accuracy while significantly improving the computational efficiency of the algorithm.

[0024] The improved optical flow model function is as follows: ; Where L is the loss function, is the gray value of the i-th frame image, T is the initial frame The extracted sub-region containing the target features, is a function of image parameterization, x is a matrix including the horizontal and vertical coordinates of the image, and v is the pixel velocity of the target area in the image.

[0025] By simplifying the motion model, the matching speed and efficiency can be improved. It is instead composed of two independent speed variables.

[0026] ; At this time, the Jacobian matrix will change to: ; At this time, the Hessian matrix changes to: ; To further speed up the algorithm, an adaptive step size is introduced to speed up the iteration speed, and a decay coefficient is added Control the acquisition of historical information to prevent the adaptive step size from accelerating too quickly, which ultimately leads to the algorithm not converging.

[0027] ; Where: is the cumulative square acceleration; is the historical information preservation degree, and its value is (0,1).

[0028] ; Update speed: ; Where: is the acceleration step length; is the total step length; It is a hyperparameter, which is given manually. First, it prevents the denominator in the calculation formula of the adaptive step size a from being zero, resulting in an undefined adaptive step size a. Second, it controls the convergence speed of the adaptive step size by updating the calculation formula of the speed V.

[0029] Repeat the iteration until v converges, then the point is determined to be the exact position of the tracking point (or target point, corner point) in this frame, and the search for the target position of the next frame begins with this point as the initial coordinate.

[0030] Repeat this cycle until the end of the video, and obtain the pixel coordinate value of the circuit breaker moving link in each frame of the left and right cameras. Then, calculate the coordinates of the moving link in the real world through the principle of binocular camera. The steps to calculate the real coordinates are as follows: refer to Figure 2 In an ideal state, two high-speed cameras are placed in parallel, and the phase planes are at the same height. For a certain point P in the shooting space, the pixel positions of the two camera images corresponding to the measurement point P are respectively and .

[0031] The conversion formula between the pixel coordinates of the camera calibration corner points and the world coordinates is: ; ; Among them, the known perspective camera model, that is, the camera's intrinsic and extrinsic matrix and , and are the Z coordinates of point P in the left and right camera coordinate systems respectively, and the real coordinates of the target point are are the pixel coordinates of the target point in the two cameras.

[0032] The actual coordinate value of the target point can be obtained by solving the above formula , and finally characterizes the three-dimensional motion trajectory of the circuit breaker.

[0033] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them.

Claims

1. A circuit breaker action characteristic measurement method based on binocular vision and improved optical flow method, characterized in that: The method is: Use dual-position cameras to obtain video data of the circuit breaker opening and closing process and detect the target point in the first frame; Find the corresponding feature points in the image based on the dual-position camera; An improved optical flow method is used to track the pixel coordinate changes of feature points in image sequences; Calculate the true three-dimensional coordinates of the feature points in the image sequence and establish a three-dimensional travel curve.

2. The circuit breaker action characteristic measurement method based on binocular vision and improved optical flow method according to claim 1, characterized in that: The method of using the improved optical flow method to track the pixel coordinate changes of feature points in the image sequence is: Establish an improved optical flow model function; Simplify the image parameterization function in the optical flow model function to obtain the Jacobian matrix and Hessian matrix; Introduce adaptive step size to accelerate iteration and add attenuation coefficient Control access to historical information; Repeat the iteration until convergence, then determine that the point is the exact position of the tracking point in this frame, and start searching for the target position of the next frame with this point as the initial coordinate; Repeat the above steps until the last frame.

3. The circuit breaker action characteristic measurement method based on binocular vision and improved optical flow method according to claim 2 is characterized in that: The improved optical flow model function is: ； Where L is the loss function, is the gray value of the i-th frame image, T is the initial frame The extracted sub-region containing the target features, is a function of image parameterization, x is a matrix including the horizontal and vertical coordinates of the image, and v is the pixel velocity of the target area in the image.

4. The circuit breaker action characteristic measurement method based on binocular vision and improved optical flow method according to claim 2, characterized in that: Function Simplified to , get the Jacobian matrix and the Hessian matrix .

5. The circuit breaker action characteristic measurement method based on binocular vision and improved optical flow method according to claim 2, characterized in that: Attenuation coefficient ; In the above formula, is the cumulative square acceleration; is the historical information preservation degree, the value is (0,1); Acceleration step length ; In the above formula, is the acceleration step length; is the total step length; is a hyperparameter.

6. The circuit breaker action characteristic measurement method based on binocular vision and improved optical flow method according to claim 5, characterized in that: Repeat until convergence, .

7. The circuit breaker action characteristic measurement method based on binocular vision and improved optical flow method according to claim 1, characterized in that: Define the real coordinate value of the target point , By solving the following formula: ； ； In the above formula, and is the pixel coordinate of the target point P in the two cameras; The conversion formula between pixel coordinates and world coordinates in the two cameras is: ； ； In the above formula, and is the camera’s intrinsic and extrinsic matrix, and are the Z coordinates of point P in the two camera coordinate systems respectively.

Images