Power transmission channel abnormal target position estimation method
Through the dual-camera visual deviation algorithm, the spatial position estimation of abnormal targets of transmission lines is achieved on edge intelligent terminals, which solves the problem of high data transmission and calculation pressure in traditional cloud recognition mode, and realizes lightweight and fast threat target recognition and early warning. Supports software and hardware upgrades of the original equipment to maintain efficient video processing.
Patent Information
- Application Number
- CN202411160084.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-22
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-08-22
AI Technical Summary
The traditional abnormal object detection mode based on cloud recognition is difficult to adapt to the wide installation range of transmission lines, resulting in excessive pressure on the transmission and computing storage of massive images and video data. The existing target positioning and dynamic recognition solutions increase the cost and computing burden. , it is difficult to effectively judge the threat movement trend of abnormal targets.
The dual-camera visual deviation algorithm is used to realize the spatial position estimation of abnormal targets on edge intelligent terminals. The position is determined through the image difference ratio and the camera distance, a cylindrical coordinate system is established, and the pixel displacement ratio is used for precise identification and estimation, which supports rapid discovery and early warning. , only a small amount of software and hardware upgrades are required, and the video processing frame rate is not affected.
It realizes rapid and accurate spatial position estimation of abnormal targets without increasing computing power and cost, supports the discovery and early warning of threat targets, saves communication and cloud resources, and maintains the efficiency of video information processing.
Smart Images

Figure CN119090961A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical fields of power transmission line abnormality monitoring, edge computing, and the like, and in particular to a method for estimating an abnormal target position in a power transmission channel. Background Art
[0002] The safe and reliable operation of transmission lines is the basis for the stable operation of power systems. Transmission lines are mainly distributed in the wild and are often threatened by abnormal targets such as foreign objects, tower cranes, and construction vehicles. In order to better deal with the impact of abnormal targets on transmission lines and reduce the occurrence of power outages, the State Grid Corporation has orderly promoted the construction of smart transmission lines in recent years. By installing various monitoring equipment on the tower, the surrounding environment of the transmission line can be monitored and monitored regularly. However, the traditional abnormal target detection mode based on cloud recognition is difficult to adapt to the needs of the power grid. This is mainly because the transmission line has a wide range of installation. If all the collected massive images and video data are uploaded to the cloud, it will not only bring huge transmission pressure to the communication link, but also bring heavy computing and storage burden to the cloud. In order to solve this problem, some scholars have proposed a solution based on edge intelligence technology, that is, embedding abnormal target detection models on the edge intelligent terminals and intelligent visualization devices of the transmission line to realize on-site rapid detection and identification of abnormal targets of the transmission line, thereby avoiding the flow and migration of massive images and videos, and effectively saving communication resources and cloud computing and storage resources.
[0003] With the computing power support of edge intelligent terminals, it is now possible to achieve real-time abnormal target recognition around power transmission lines at more than 33 frames per second (the minimum frame rate requirement for real-time monitoring of power transmission lines). However, identifying targets alone is not enough to correctly judge threats, or it is difficult to effectively distinguish whether abnormal targets have a movement trend that threatens transmission lines. Therefore, it is urgent to overcome the spatial coordinate estimation method for abnormal targets.
[0004] However, if the existing target positioning and dynamic identification solutions are further added on the basis of the current transmission line monitoring system, the cost and computing power burden will be greatly increased. For example, the existing technologies of lidar 3D point cloud scanning, binocular depth estimation, etc., will require a lot of additional costs, and the computing power required to implement the corresponding identification algorithm also exceeds the computing power level of the current edge intelligent terminals configured for transmission lines. Summary of the invention
[0005] Therefore, in view of the actual needs of abnormal monitoring of power transmission lines, and the consideration of cost and difficulty of transformation, the present invention proposes a transmission channel abnormal target position estimation scheme. The proposed dual-camera based visual deviation algorithm can achieve rapid estimation of the spatial position of abnormal targets without consuming too much computing power to reconstruct the on-site three-dimensional model, and support the discovery, alarm and even early warning of threatening abnormal targets. It can be built on the basis of the original target recognition algorithm and hardware system, and only a small amount of software and hardware upgrades and adjustments are needed to re-enable the original equipment; at the same time, it has the characteristics of being lightweight and will not affect the edge intelligent gateway's processing frame rate for video information.
[0006] The present invention specifically adopts the following technical solutions: A method for estimating the position of abnormal targets in power transmission channels: Two cameras of the same model are installed on two adjacent transmission towers. The dual cameras will be networked with the same edge intelligent computing terminal and perform calculations locally to estimate the location of abnormal targets in the transmission channel: The ratio of the difference between the images of the abnormal target collected by the two cameras is used. When the distance between the two transmission towers is known, the location of the abnormal target is determined by the unique correspondence between the ratio and the distance to the camera.
[0007] Furthermore, the geometric midpoint of the abnormal target identification frame is used to represent the abnormal target position, and the change of the abnormal target position between two frames collected by two cameras is used to realize the displacement estimation of the abnormal target in the direction of the transmission channel by using the pixel displacement ratio.
[0008] Furthermore, two cameras are set at the plane geometric center of all line mounting points, and it is ensured that the lowest point of the line sag should appear at the near end of the field of view.
[0009] Among them, the camera and the edge intelligent computing terminal are standard device configurations for existing power transmission network monitoring. The improvement of the device in the present invention is that the setting position of the camera is readjusted to facilitate the construction of a corresponding cylindrical field of view, so that abnormal targets can be located using a streamlined method without further adding or replacing equipment.
[0010] In terms of identification methods, since the identification of target types is already a routine configuration for existing transmission line monitoring, this solution converts the problem of identifying abnormal targets into pixel displacement of the midpoint of the detection frame; and cleverly uses the symmetry by utilizing the difference in images captured by the two cameras to achieve more accurate identification, and only involves simple table lookup or linear calculation, which basically does not increase the burden on the edge intelligent computing terminal, thereby achieving localized processing.
[0011] Furthermore, a cylindrical coordinate system is established, with the line connecting the two cameras as the z-axis and the origin of the z-axis set at one of the cameras; the relative position coordinates of the abnormal target are set as ( ),in is the radial distance in the cylindrical coordinate system, is the azimuth, is the coordinate on the z-axis; When an abnormal target enters the field of view of one of the cameras, the edge intelligent computing terminal first identifies it and waits for the target with the same name to appear in the field of view of the other camera; after confirming that the target appears in two opposite fields of view, the edge intelligent computing terminal takes the current frame image and the next frame image of the dual cameras to analyze the movement of the abnormal target.
[0012] The consideration of this processing is to construct a "visual fence" through a cylindrical coordinate system. Only abnormal targets entering the fence area are considered to be a possible threat to the transmission line, and further monitoring is initiated, thereby saving computing power.
[0013] Furthermore, by pre-establishing the pixel displacement ratio and the cross-sectional position of the cylindrical coordinate system in the direction of the transmission channel The one-to-one correspondence between the current frame image and the next frame image is realized by the pixel displacement ratio of The estimation of the pixel points on the image is further realized by using the intrinsic parameter matrix of the camera. The one-to-one correspondence of the coordinate points on the plane is used to obtain the spatial coordinates of the abnormal target ( ).
[0014] Furthermore, when the abnormal target continues to appear in the field of view of the two cameras, the displacement of the abnormal target in the direction of the transmission channel is estimated by utilizing the change in the position of the center point of the two sets of images between two frames and using the pixel displacement ratio.
[0015] Furthermore, when estimating the positions of multiple abnormal targets: For multiple different abnormal targets, the dual-camera recognition targets are grouped and paired by identifying the names of the targets; For multiple identical abnormal targets, based on the multi-frame images of the dual cameras, the recognition frames belonging to the same abnormal target should have the same inter-frame displacement ratio, and the same abnormal target frames are matched by the ratio.
[0016] Furthermore, if the abnormal target stops moving after entering the horizons on both sides, the ratios of each type will not be calculated and updated, and only the previous calculation results will be used. If the abnormal target moves out of one side of the horizon, the position estimation will be stopped; if the abnormal target only appears in the horizon on one side, no position estimation will be performed.
[0017] This is because the different angle shapes of abnormal targets may be uncertain and unknown. Only by using the movement of the center of the recognition frame of this solution and calculating the proportion can a relatively high accuracy rate be achieved. In addition, targets that temporarily stop moving can generally be considered to pose no further risks (an alarm can be triggered when entering the dual field of view), so there is no need to invest computing resources for the time being.
[0018] Furthermore, the frame rate of the acquired images during the abnormal target position estimation calculation process is adjusted to match the hardware processing capability and / or ensure that the obvious displacement of the abnormal target is captured.
[0019] This adjustability is designed to match existing target acquisition and recognition solutions and save computing resources as much as possible.
[0020] And, a corresponding device for estimating the position of an abnormal target in a power transmission channel, comprising: two cameras of the same model and an edge intelligent computing terminal connected in a network; Two cameras are set up on two adjacent transmission towers facing each other. The ratio of the difference in images of the abnormal target captured by the two cameras is used. When the distance between the two transmission towers is known, the location of the abnormal target is determined by the unique correspondence between the ratio and the distance to the camera.
[0021] Compared with the existing technology, the present invention and its preferred solution can quickly estimate the spatial position of abnormal targets in the power transmission channel, and support the discovery, alarm and even early warning of threatening abnormal targets. It is based on the original target recognition algorithm and hardware system, and only a small amount of software and hardware upgrades are needed to re-enable the original equipment; at the same time, it is lightweight and will not affect the processing frame rate of the edge intelligent gateway for video information. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments: Figure 1 This is a schematic diagram of the dual camera settings according to an embodiment of the present invention; Figure 2 Schematic diagram of the position estimation principle of an embodiment of the present invention; wherein: a) a spatial view; b) a top view. DETAILED DESCRIPTION
[0023] Hereinafter, the specific embodiments of the present application will be described in detail with reference to the accompanying drawings. According to these detailed descriptions, those skilled in the art can clearly understand the present application and can implement the present application. Without violating the principles of the present application, the features in the various embodiments can be combined to obtain new implementations, or certain features in certain embodiments can be replaced to obtain other preferred implementations.
[0024] In order to make the features and advantages of this patent more obvious and easy to understand, the following embodiments are specifically described in detail with the accompanying drawings as follows: 1. Dual camera setup There are various types of transmission towers. This invention takes a group of cat-head type transmission towers as an example to illustrate a general dual-camera setting principle. Figure 1 As shown. In order to effectively associate the camera parallax information with the spatial position of the abnormal target, it is necessary to ensure that the two cameras are set up facing each other. At the same time, the general requirements for the camera setting position for power transmission channel monitoring should be met, that is, the lowest point of the line sag should appear at the near end of the field of view. If conditions permit, the camera should be set at the plane geometric center of all line mounting points, such as Figure 1 As shown by the green dot. This can more comprehensively ensure the safety of the line and realize the timely detection and early warning of abnormal targets approaching from all directions. The dual cameras will form a network with the same edge intelligent computing terminal. At this time, the intelligent terminal can use the algorithm of the present invention and its own computing power to realize the correlation analysis of the visual information (video) sent back by the dual cameras.
[0025] 2. Position estimation principle and algorithm In order to effectively evaluate the spatial position of abnormal targets and determine the threat level of abnormal targets to the power transmission channel, a cylindrical coordinate system is first established. The line connecting the two opposite cameras is the z-axis, and the origin of the z-axis is at the camera of the n# tower. The subsequent algorithm will use this as a reference to realize the relative position coordinates of the abnormal targets ( ), where is the radial distance, is the azimuth, is the coordinate on the z-axis, such as Figure 2 shown.
[0026] When an abnormal target enters the field of view of one camera, the edge intelligent computing terminal will first complete the identification and wait for the same target to appear in the field of view of the other camera. After confirming that the target appears in two opposite fields of view, the edge intelligent terminal will take the current frame image and the next frame image of the dual cameras to analyze the movement of the abnormal target.
[0027] The specific analysis method and algorithm principles involved are: a) Taking the geometric midpoint of the abnormal target recognition frame to represent the abnormal target position can effectively eliminate the influence of the displacement of the abnormal target in the z-axis direction on the accuracy of spatial position estimation; b) Using the change in the position of the center point of the two sets of images between two frames, the displacement of the target in the two sides of the field of view is calculated respectively. The displacement distance is measured in the number of pixels, that is, pixel displacement. It should be noted that if the original video processing frame rate is high, it is necessary to consider selecting images for analysis every other frame so that the abnormal target has enough time to complete the obvious displacement, thereby ensuring the accuracy of the estimation algorithm; c) In most cases, due to the different distances between the abnormal target and the cameras on both sides, the displacement performance (pixel displacement) of the abnormal target in the two fields of view will be different. However, in fact, the displacement of the abnormal target on the normal section of the z-axis is consistent. Therefore, the present invention establishes a relationship between the pixel displacement ratio and the cross-sectional position. The one-to-one correspondence between the pixel displacement ratio can be effectively used to achieve The specific implementation methods include but are not limited to table building and linear fitting. The input of the algorithm should be the pixel displacement ratio, and the output should be Get the value. After that, the intrinsic parameter matrix of the camera can be used to realize the pixel points on the image and The coordinate points on the plane correspond one to one, that is, the spatial coordinates of the abnormal target are obtained ( ).
[0028] d) Repeat the above steps to achieve dynamic tracking of the abnormal target location.
[0029] The above is the principle for estimating the spatial position of a single abnormal target.
[0030] In order to better handle the situation where multiple targets enter the field of view at the same time, the following processing logic may be added in this embodiment: 1) Distinguishing between multiple abnormal targets: When an abnormal target enters two fields of view successively, the names of the targets on both sides will be compared first. After ensuring that the names appear in pairs, the recognition frames will be grouped based on the names.
[0031] 2) Distinguishing multiple identical abnormal targets: When the third target with the same name appears in the field of view, first release the push of the previous estimation results, and then wait for the appearance of the fourth target with the same name. If the fourth target appears, calculate the displacement length of the geometric center of the four target recognition frames between the three frames and calculate the ratio of the displacement of each center point between the first and second frames to the displacement between the second and third frames, that is, the displacement ratio between consecutive frames. At this time, the recognition frames belonging to the same abnormal target should have the same property of inter-frame displacement ratio, so by comparing the four ratios, effective matching of the recognition frames on both sides can be achieved.
[0032] It should be noted that if the abnormal target stops moving after entering the horizons on both sides, the ratios will not be calculated and updated, and only the previous calculation results will be used. If the target moves out of one side of the horizon, the position estimation will be stopped; in addition, if the target only appears in the horizon on one side, no position estimation will be performed.
[0033] The above is a preferred embodiment of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.
[0034] This patent is not limited to the above-mentioned optimal implementation mode. Anyone can derive various other forms of a method for estimating the abnormal target position in a transmission channel under the inspiration of this patent. All equal changes and modifications made according to the scope of the patent application of the present invention should be covered by this patent.
Claims
1. A method for estimating the position of an abnormal target in a power transmission channel, characterized in that: Two cameras of the same model are installed on two adjacent transmission towers. The dual cameras will be networked with the same edge intelligent computing terminal and perform calculations locally to estimate the location of abnormal targets in the transmission channel: The ratio of the difference between the images of the abnormal target collected by the two cameras is used. When the distance between the two transmission towers is known, the location of the abnormal target is determined by the unique correspondence between the ratio and the distance to the camera.
2. The method for estimating abnormal target position in a power transmission channel according to claim 1, characterized in that: The geometric midpoint of the abnormal target identification frame is used to represent the abnormal target position. The change of the abnormal target position between two frames collected by two cameras is used, and the pixel displacement ratio is used to realize the displacement estimation of the abnormal target in the direction of the transmission channel.
3. The method for estimating abnormal target position in a power transmission channel according to claim 1, characterized in that: Set the two cameras at the plane geometric center of all line mounting points, and ensure that the lowest point of the line sag should appear at the near end of the field of view.
4. The method for estimating abnormal target position in a power transmission channel according to claim 2, characterized in that: Establish a cylindrical coordinate system with the line connecting the two cameras as the z-axis and the origin of the z-axis as one of the cameras; let the relative position coordinates of the abnormal target be ( ),in is the radial distance in the cylindrical coordinate system, is the azimuth, is the coordinate on the z-axis; When an abnormal target enters the field of view of one of the cameras, the edge intelligent computing terminal first identifies it and waits for the target with the same name to appear in the field of view of the other camera; after confirming that the target appears in two opposite fields of view, the edge intelligent computing terminal takes the current frame image and the next frame image of the dual cameras to analyze the movement of the abnormal target.
5. A method for estimating abnormal target position in a power transmission channel according to claim 4, characterized in that: By pre-establishing the pixel displacement ratio and the cross-sectional position of the cylindrical coordinate system in the direction of the transmission channel The one-to-one correspondence between the current frame image and the next frame image is realized by the pixel displacement ratio of The estimation of the pixel points on the image is further realized by using the intrinsic parameter matrix of the camera. The one-to-one correspondence of the coordinate points on the plane is used to obtain the spatial coordinates of the abnormal target ( ).
6. A method for estimating abnormal target position in a power transmission channel according to claim 4, characterized in that: When the abnormal target continues to appear in the field of view of the two cameras, the displacement of the abnormal target in the direction of the transmission channel is estimated by utilizing the change in the position of the center point of the two sets of images between the two frames and using the pixel displacement ratio.
7. The method for estimating abnormal target position in a power transmission channel according to claim 2, characterized in that: When estimating the positions of multiple abnormal targets: For multiple different abnormal targets, the dual-camera recognition targets are grouped and paired by identifying the names of the targets; For multiple identical abnormal targets, based on the multi-frame images of the dual cameras, the recognition frames belonging to the same abnormal target should have the same inter-frame displacement ratio, and the same abnormal target frames are matched by the ratio.
8. The method for estimating abnormal target position in a power transmission channel according to claim 2, characterized in that: If the abnormal target stops moving after entering the horizons on both sides, the ratios will not be calculated and updated, and only the previous calculation results will be used. If the abnormal target moves out of one side of the horizon, the position estimation will be stopped; if the abnormal target only appears in the horizon on one side, no position estimation will be performed.
9. The method for estimating abnormal target position in a power transmission channel according to claim 2, characterized in that: The frame rate of the acquired images during the abnormal target position estimation calculation process is adjusted to match the hardware processing capability and / or to ensure that the obvious displacement of the abnormal target is captured.
10. A device for estimating abnormal target position in a power transmission channel, characterized in that: Includes the following networking components: two cameras of the same model and an edge intelligent computing terminal; Two cameras are set up opposite each other on two adjacent transmission towers. The ratio of the difference in images of the abnormal target captured by the two cameras is used. When the distance between the two transmission towers is known, the location of the abnormal target is determined by the ratio and the specific unique correspondence with the cameras.
Citation Information
Patent Citations
Tracking system based on binocular camera shooting
CN101344965A
Device and method for detecting sag and limited distance of power transmission line
CN110953993A
Pole tower monitoring method and terminal based on digital image processing
CN114119746A
Power transmission line foreign matter and engineering vehicle monitoring system based on side cloud cooperation
CN114973132A
Vehicle speed intelligent measurement method based on binocular stereo vision system
US20220405947A1
