A Three-Dimensional Vibration Detection Method for Transformers Based on Binocular Ranging and Optical Flow

Abstract

A method for three-dimensional vibration detection of transformers based on binocular ranging and optical flow is disclosed. In the method, a binocular camera acquires real-time video signals of two transformers and calculates the transformer depth information; extracts feature points of the transformer from one video stream; calculates two-dimensional vibration vectors of all the feature points for feature point tracking; calculates three-dimensional vibration vectors of the feature points based on the depth information; calculates the vibration frequency of all the feature points; and visualizes the vibration state of the transformer based on the vibration vectors and vibration frequencies.

Description

Technical Field

[0001] This invention belongs to the field of transformer three-dimensional vibration detection technology, and particularly relates to a transformer three-dimensional vibration detection method based on binocular ranging and optical flow method. Background Technology

[0002] Transformers are one of the key pieces of equipment in power systems, and their operating status has a significant impact on the safe and stable operation of the power grid. With the advancement of intelligent technologies, condition-based maintenance is gradually replacing traditional periodic maintenance and post-accident repair methods for transformer maintenance. Currently, transformer condition assessment methods are mainly divided into two categories: online monitoring and offline detection. Online monitoring allows transformers to operate without interruption, saving manpower and resources, and has significant advantages. Commonly used online monitoring methods for power transformers include: low-voltage pulse method, frequency response analysis method, gas chromatography analysis of dissolved gases in transformer oil, partial discharge online monitoring method, and vibration analysis method. Among these, a major advantage of vibration analysis method is that the detection system does not have any electrical connection with the transformer under test, and will not affect the normal operation of the power grid, fully ensuring the safety of online monitoring.

[0003] Vibration signal detection is a crucial prerequisite for analyzing and evaluating the operating status of transformers and diagnosing faults using vibration analysis methods. Vibration signal detection methods can be broadly categorized into contact and non-contact methods. Traditional contact vibration measurement primarily employs on-site sensor installation, which, due to the need for point-by-point deployment, has a limited measurement range and several drawbacks. Currently, commonly used non-contact vibration measurement methods are mainly divided into laser vibration measurement and image / video-based visual vibration measurement methods. While laser vibration measurement offers advantages such as high accuracy and sensitivity, long measurement distance, and high measurement frequency, the related equipment is very expensive, and it requires a high level of expertise from operators, significantly hindering its widespread adoption and application. Visual vibration measurement, as an emerging vibration measurement method, has attracted widespread attention from scholars both domestically and internationally. Video image monitoring is a non-contact monitoring method capable of both static measurements such as displacement and strain, as well as dynamic characteristic measurements. It offers advantages such as ease of operation, non-contact nature, non-destructive testing, no added mass, and the ability to achieve long-distance, large-scale, multi-point monitoring. However, it also has drawbacks, including the need for target placement, high requirements for monitoring equipment, and strict requirements for ambient light and background conditions.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of the present invention, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a method for three-dimensional vibration detection of transformers based on binocular ranging and optical flow. To achieve the above objective, this invention provides the following technical solution:

[0006] The present invention provides a method for three-dimensional vibration detection of transformers based on binocular ranging and optical flow, comprising:

[0007] Step S100: The binocular camera acquires real-time video signals from the dual-channel transformer and calculates the transformer depth information;

[0008] Step S200: Extract feature points of the real-time video signal of the transformer in a single video stream;

[0009] Step S300: Calculate the two-dimensional vibration vector of all the feature points, and obtain the three-dimensional coordinates of the feature points by combining the two-dimensional vibration vector with the depth information. Further obtain the three-dimensional vibration vector based on the three-dimensional coordinates.

[0010] Step S400: Calculate the vibration frequency of all the aforementioned feature points;

[0011] Step S500: Visualize the vibration state of the transformer based on the three-dimensional vibration vector and vibration frequency. In the aforementioned three-dimensional vibration detection method for transformers based on binocular ranging and optical flow, the vibration vector of all feature points is calculated using the pyramid LK optical flow method. The pyramid LK optical flow method is expressed as:

[0012] I x (q1)V x +I y (q1)V y =-I t (q1)

[0013] I x (q2)V x +I y (q2)V y =-I t (q2)

[0014]

[0015] I x (q n V x +I y (q n V y =-I t (q n ),

[0016] Where q1, q2, ..., q n I represents the pixels contained within a small window, which delineates an N×N region around a feature point. x (q i ), I y (q i ) and It (q i ) represent pixels q i The partial derivatives of the light intensity I with respect to the x, y, and t directions, V x V represents the speed at which a pixel moves in the x-direction. y The speed at which the pixel moves in the y-direction.

[0017] In the aforementioned method for three-dimensional vibration detection of transformers based on binocular ranging and optical flow, in step S200, the Shi-Tomasi algorithm is used to extract feature points of the transformer.

[0018] In the aforementioned method for detecting three-dimensional vibration of a transformer based on binocular ranging and optical flow, in step S400, a two-dimensional vibration vector is extracted frame by frame for each feature point, and a three-dimensional coordinate sequence is obtained by combining it with depth information and performing a fast Fourier transform to obtain the vibration frequency of each feature point.

[0019] In the above technical solution, the present invention provides a transformer three-dimensional vibration detection method based on binocular ranging and optical flow, which has the following beneficial effects: The transformer three-dimensional vibration detection method based on binocular ranging and optical flow uses the Shi-Tomasi method to extract feature points to track transformer vibration, thereby improving the anti-interference ability of environmental factors, light, etc.; it uses a binocular camera to generate actual transformer three-dimensional vibration data, improving the authenticity and reliability of the analysis data; and it visualizes the three-dimensional vibration data frame by frame, enabling maintenance personnel to intuitively analyze and judge the operating status of the transformer. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0021] Figure 1 This is a flowchart illustrating an embodiment of the transformer three-dimensional vibration detection method based on binocular ranging and optical flow in this invention.

[0022] Figure 2 This is a schematic diagram of the binocular ranging principle of an embodiment of the three-dimensional vibration detection method for transformers based on binocular ranging and optical flow in this invention.

[0023] Figure 3 This is a schematic diagram illustrating an embodiment of the transformer three-dimensional vibration detection method based on binocular ranging and optical flow in this invention, which uses the pyramid method to track feature points from the highest level of the pyramid to the lowest level.

[0024] Figure 4 This is a schematic diagram of the visualization of transformer vibration state in the three-dimensional vibration detection method for transformers based on binocular ranging and optical flow method in this invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Therefore, the following is an explanation of the figures. Figures 1 to 4 The detailed description of the embodiments of the present invention provided herein is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0027] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0028] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0029] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0030] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0031] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0032] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings. Figure 1 As shown, a method for detecting three-dimensional vibration of a transformer based on binocular ranging and optical flow includes,

[0033] Step S100: The binocular camera acquires real-time video signals from the dual-channel transformer and calculates the transformer depth information;

[0034] Step S200: Extract feature points of the real-time video signal of the transformer in a single video stream;

[0035] Step S300: Calculate the two-dimensional vibration vector of all the feature points, and obtain the three-dimensional coordinates of the feature points by combining the two-dimensional vibration vector with the depth information. Further obtain the three-dimensional vibration vector based on the three-dimensional coordinates.

[0036] Step S400: Calculate the vibration frequency of all the aforementioned feature points;

[0037] Step S500: Visualize the vibration state of the transformer based on the three-dimensional vibration vector and vibration frequency.

[0038] In a preferred embodiment of the transformer three-dimensional vibration detection method based on binocular ranging and optical flow, the vibration vector of all feature points is calculated using the pyramid LK optical flow method, which is expressed as:

[0039] I x (q1)V x +Iy (q1)V y =-I t (q1)

[0040] I x (q2)V x +I y (q2)V y =-I t (q2)

[0041]

[0042] I x (q n V x +I y (q n V y =-I t (q n ),

[0043] Where q1, q2, ..., q n I represents the pixels contained within a small window, which delineates an N×N region around a feature point. x (q i ), I y (q i ) and I t (q i ) represent pixels q i The partial derivatives of the light intensity I with respect to the x, y, and t directions, V x V represents the speed at which a pixel moves in the x-direction. y The speed at which the pixel moves in the y-direction.

[0044] In a preferred embodiment of the transformer three-dimensional vibration detection method based on binocular ranging and optical flow, in step S200, the Shi-Tomasi algorithm is used to extract the feature points of the transformer.

[0045] In a preferred embodiment of the transformer three-dimensional vibration detection method based on binocular ranging and optical flow, in step S400, a two-dimensional vibration vector is extracted frame by frame for each feature point, and a three-dimensional coordinate sequence is obtained by combining the depth information and performing a fast Fourier transform to obtain the vibration frequency of each feature point.

[0046] In one embodiment, the three-dimensional vibration detection method for transformers includes the following steps:

[0047] S100: Load the binocular camera and calculate the transformer depth information;

[0048] In this step, such as Figure 2As shown, P is the vertex coordinate of the transformer, and its coordinates are related to the optical center C of the left camera. l and the optical center C of the right camera r The intersection point P′ of the line connecting the left and right camera planes. L 、P′ R That is, the projection point, which can be known from similar triangles:

[0049]

[0050] Where b is the distance between the left and right cameras, f is the focal length of the left and right cameras, and X L X R Let d be the distance from the center of the image position of the same pixel on the left and right cameras, and d be the parallax. The depth of the transformer.

[0051] S200: Extract feature points of the transformer;

[0052] S300: Calculates the vibration vector of all feature points;

[0053] S400: Calculates the vibration frequencies of all characteristic points;

[0054] S500: Visualize the vibration state of the transformer based on the vibration vector and vibration frequency of the feature points.

[0055] In another embodiment, the Shi-Tomasi algorithm is used to extract feature points of the transformer.

[0056] In this step, the pixels corresponding to the local maximum values ​​in the grayscale gradient are the corner points. A fixed-size window is moved slightly in any direction across a certain location on the image. If the grayscale values ​​(on the gradient map) within the window change significantly, then a corner point exists in the region containing that window. Corner detection includes the following three steps:

[0057] Step 1: Calculate the change in pixel values ​​within the window as the window (a small image fragment) moves simultaneously in both the x and y directions;

[0058] Step 2: Calculate the score (corner response function) R for each window based on the change in pixel values;

[0059] Step 3: Perform threshold processing on the function. If R is greater than a set threshold, it means that the window corresponds to a corner feature.

[0060] In another embodiment, the vibration vectors of all feature points are calculated using the pyramid LK optical flow method.

[0061] In this embodiment, the LK optical flow method adds the assumption of "spatial consistency" to the two basic assumptions of the original optical flow method, that is, adjacent pixels within a certain region have approximately the same movement. Based on this assumption, an N×N window can be drawn around the feature point, and it is assumed that all pixels within the window have the same movement. The LK optical flow method can be described by the following formula:

[0062]

[0063] In the formula, q1, q2, ..., q n I represents the number of pixels contained within the small window. x (q i ), I y (q i ) and I t (q i ) represent pixels q i The partial derivatives of the light intensity I in the x, y, and t directions.

[0064] Furthermore, formula (2) can be expressed in the following matrix form:

[0065]

[0066] In the basic optical flow method, because there is only one constraint equation, it is impossible to solve for two unknowns. However, in the LK optical flow method, there are n equations (n>2), resulting in an overdetermined system of equations. Solving this system of equations using the least squares method yields:

[0067]

[0068] in, This represents the speed of movement in the x-direction; This represents the speed of movement in the y-direction.

[0069] To prevent large errors that may occur when the object's motion amplitude is large, the pyramid method is used to improve the LK optical flow method, such as... Figure 3 As shown, the pyramid LK optical flow method downsamples the image, scaling it to various sizes. This allows for the reduction of large displacements within the higher-level pyramid images. The LK method is then used to estimate the optical flow between images 1 and 2, yielding a relatively accurate optical flow vector at that scale. Then, as the solution is applied layer by layer from the top, interpolation and upsampling are used to proportionally amplify the higher-level vectors as initial guides for the next layer. At this point, there will be an error between the target position indicated by the amplified optical flow vector from the higher levels and the actual target position in the current layer. However, this error usually corresponds to the scale of small movements. Therefore, the optical flow vector of the current layer can be calculated based on this error. By iteratively descending to the bottom original image, a relatively accurate optical flow vector under large-amplitude movements can be obtained.

[0070] In another embodiment, the vibration frequency of each feature point is obtained by extracting its three-dimensional coordinate sequence frame by frame and performing a fast Fourier transform.

[0071] In this embodiment, each feature point is tracked frame by frame in the video sequence. When a specified number of frames is reached, the three-dimensional coordinates of each feature point are stored as a discrete sequence according to the frame order. A Fast Fourier Transform (FFT) is performed on the sequence to obtain the vibration spectrum of each feature point, and the frequency with the largest amplitude is taken as the instantaneous vibration frequency of that feature point. Simultaneously, the displacement vector of the feature point between the current frame and the previous frame is taken as the vibration vector. The same calculation is performed on all feature points sequentially to obtain the instantaneous vibration frequencies of all feature points in the current frame. To calculate the vibration data for the next frame, the data from the first frame in the three-dimensional coordinate discrete sequence is removed, and the three-dimensional coordinates of the next frame are added. The vibration frequency and vibration vector are calculated similarly.

[0072] Figure 4 The brightness of feature points in a single frame varies with the vibration frequency. Feature points are displayed in red, and their brightness is adjusted based on their vibration frequency and vector magnitude. This is then displayed frame-by-frame in a video sequence, allowing observation of the significant brightness changes of the transformer as it vibrates.

[0073] Finally, it should be noted that the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0074] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for three-dimensional vibration detection of transformers based on binocular ranging and optical flow, characterized in that, It includes the following steps: Step S100: The binocular camera acquires real-time video signals from the dual-channel transformer and calculates the transformer depth information. These are the vertex coordinates of the transformer, and their coordinates are the optical center of the left camera. and the optical center of the right camera The intersection of the line connecting the left and right camera planes , That is, the projection point, which can be known from similar triangles: （1） in, The distance between the left and right cameras. The focal lengths of the left and right cameras. , This refers to the distance from the center of the image position of the same pixel on both the left and right cameras. For parallax, The depth of the transformer; Step S200: Extract feature points of the real-time video signal of the transformer in a single video stream. Use the Shi-Tomasi algorithm to extract these feature points. The pixel corresponding to the local maximum value in the grayscale gradient is the corner point. Use a fixed-size window to make tiny slides in any direction at a certain position on the image. First step: When the window simultaneously moves towards... and When moving in two directions, calculate the change in pixel values ​​inside the window; the second step: calculate the corner response function R corresponding to each window based on the change in pixel values; perform thresholding on the corner response function R, if R is greater than a set threshold, it means that the window corresponds to a corner feature; Step S300: Calculate the two-dimensional vibration vector of all the feature points, and obtain the three-dimensional coordinates of the feature points by combining the two-dimensional vibration vector with the depth information. Further obtain the three-dimensional vibration vector based on the three-dimensional coordinates. Step S400: Calculate the vibration frequency of all the feature points, track each feature point frame by frame, and when the specified number of frames is reached, store the three-dimensional coordinates of each feature point as a discrete sequence according to the frame order. Perform a fast Fourier transform on the discrete sequence to obtain the vibration spectrum of each feature point, and take the frequency with the largest amplitude as the instantaneous vibration frequency of the feature point. At the same time, take the displacement vector of the feature point between the current frame and the previous frame as the vibration vector. Perform the same calculation on all feature points in turn to obtain the instantaneous vibration frequency of all feature points in the current frame. Step S500: Visualize the vibration state of the transformer based on the three-dimensional vibration vector and vibration frequency.

2. The method for three-dimensional vibration detection of a transformer based on binocular ranging and optical flow as described in claim 1, characterized in that, The vibration vectors of all feature points are calculated using the pyramid LK optical flow method, which is expressed as follows: ， in, , , , This represents the number of pixels contained within a small window, which is a boundary drawn around a feature point. Scope , and Representing pixels light intensity exist , , Partial derivatives in the direction, For pixels in Speed ​​of movement in a certain direction; For pixels in The speed of movement in a certain direction.

3. The method for three-dimensional vibration detection of a transformer based on binocular ranging and optical flow as described in claim 1, characterized in that, In step S400, the two-dimensional vibration vector of each feature point is extracted frame by frame, and the three-dimensional coordinate sequence is obtained by combining the depth information and performing a fast Fourier transform to obtain the vibration frequency of each feature point.

Images