A method and system for identifying and warning external damage of a power transmission line

Abstract

The application discloses a kind of power transmission line external damage identification warning method and system, method is: the target object of power transmission line protection area is extracted and segmented to obtain target identification result;Different voltage grade power transmission line protection area is demarcated alarm area rule frame and is identified frame selection, combined with target identification result generates image identification warning;Target object action in a period of time is acquired, whether target object exists approach behavior is analyzed, video background extraction and interframe difference are used to generate video analysis warning;The laser point cloud model of power transmission line span and its channel range is acquired, three-dimensional scene reconstruction is carried out and is matched with two-dimensional image reconstruction and obtains target object, the shortest distance of target object and power transmission line is measured and screened out, generates three-dimensional point cloud warning;Comprehensive analysis result and execute alarm action;Shorten the investigation cost of power transmission line external damage hidden danger, improve power transmission line short time external damage identification efficiency and warning accuracy.

Description

Technical Field

[0001] The present invention belongs to the technical field of preventing power transmission lines from being damaged by external forces, and in particular relates to a method and system for identifying and warning of external force damage to power transmission lines. Background Art

[0002] With the digital transformation of power grids and the development of intelligent technologies, remote video monitoring has initially solved the problem of on-site supervision of transmission lines, greatly facilitating daily maintenance for operators. Consequently, the number of tower-mounted video equipment installed is increasing. Tower-mounted video equipment primarily captures images from pre-positioned locations. Due to the large number of devices, hundreds of thousands of images are generated daily. Operators inspect each image individually to identify any threats to the safety and stability of the transmission lines, which is time-consuming and inefficient. This also provides insufficient real-time and unpredictable monitoring of external forces, such as construction machinery. There is an urgent need to improve the accuracy of external damage identification and alarms, as well as the efficiency of video monitoring.

[0003] Existing image recognition for external damage primarily uses deep learning algorithms to detect and identify construction machinery and generate alerts. However, these methods suffer from significant background interference and a wide object detection range. To ensure no missed detections, the confidence interval must be expanded, resulting in low accuracy for external damage detection alerts and requiring maintenance personnel to inspect a wide range of abnormal images. Existing technologies primarily employ single-layer convolutional neural network recognition methods for capturing captured images. This approach works by first capturing objects within the field of view and comparing them with common construction machinery, then outputting the corresponding abnormal object for each image. However, this method has a relatively wide image recognition range. Because some construction machinery in real-world scenarios is not within a safe distance from the transmission line protection zone or is only briefly present and poses no threat of approaching dangerously close, existing single-layer convolutional neural network object detection methods lack accuracy and real-time performance in practical applications. They are unable to accurately reflect real-time hazards within the transmission line protection zone. Furthermore, existing technologies lack a multi-faceted verification method for external damage detection alerts. Summary of the Invention

[0004] The main purpose of the present invention is to overcome the shortcomings and deficiencies of the existing technology and provide a method and system for identifying and warning of external force damage to transmission lines. The method adopts a comprehensive analysis of target detection and feature extraction, video frame analysis and laser point cloud three-dimensional ranging to perform graded alarm actions on external force damage behaviors in the transmission line protection zone, thereby improving the accuracy and real-time performance of external force damage identification and alarm, thereby shortening the time spent on video monitoring and image inspection, and improving the efficiency of external force damage to transmission lines.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In one aspect, the present invention provides a method for identifying and warning of external force damage to a power transmission line, characterized in that it comprises the following steps:

[0007] Classify the target objects in the transmission line protection zone within the camera's field of view, extract and segment the camera-captured images based on target detection and convolutional neural networks, and obtain target recognition results.

[0008] Using the tower within the camera's field of view as the center line, the system demarcates alarm areas for transmission line protection zones of different voltage levels and performs identification and selection. Combined with the target recognition results, it generates image recognition alarms for target objects threatening transmission lines.

[0009] Initiate real-time video recording based on image recognition alarms, capture the target object's movement over a period of time, extract video frames containing the target object, combine target detection to analyze whether the target object is approaching the power transmission line, and generate video analysis alarms using video background extraction and inter-frame difference.

[0010] Obtain a 3D laser point cloud model of the transmission line span and its channel range, perform spatial coordinate transformation on the target within the camera's field of view, and achieve corresponding reconstruction of the 3D point cloud scene in the 2D image;

[0011] Based on the target objects and their locations in the transmission line protection zone that are matched and reconstructed in the 3D scene, the multi-directional distance measurement between the target objects and the transmission lines is performed using the principle of similar triangles. The shortest distance is selected and compared with the safe distance of the transmission line at the corresponding voltage level to generate a 3D point cloud alarm.

[0012] Comprehensively analyze image recognition alarms, video analysis alarms and three-dimensional point cloud alarms, mutually verify and detect the output of transmission line external force damage alarm information. If two or more of the three issue an alarm, the transmission line external force damage alarm action will be immediately executed.

[0013] As a preferred technical solution, the target detection adopts the YOLOv4 target detection network structure and uses the K-means algorithm to perform cluster analysis on the captured images;

[0014] The convolutional neural network is constructed based on TensorFlow and Faster-RCNN network. The convolutional neural network is trained using hierarchical weighted cross entropy as the loss function, and the Focal Loss algorithm is used to optimize the loss function. The convolutional neural network is used for two learning and training sessions to extract target features and segment images to obtain target recognition results.

[0015] As a preferred technical solution, the generating of image recognition alarm is specifically as follows:

[0016] Within the camera's field of view, the transmission line tower is used as the center line, and the safety distances of different voltage levels are used as parallel lines. The area between the two parallel lines is designated as the alarm area.

[0017] Target detection is used to identify and select transmission line towers and protection zones of different voltage levels. Combined with the target recognition results, when the target object enters the alarm area, an image recognition alarm is generated for the target object that threatens the transmission line.

[0018] As a preferred technical solution, the generation of video analysis alarms is specifically as follows:

[0019] The moving motion of the target object is verified and matched by improving the mixed Gaussian model to obtain the background image and extract the target;

[0020] The improved three-frame difference method is used to perform temporal difference between the image frame of the target object in motion and the background image to obtain the moving target pixel area;

[0021] The background image, the extracted target and the moving target pixel area are integrated, and the transmission line and tower are used as reference objects to analyze whether the target object is approaching the transmission line. If so, a video analysis alarm is generated and a warning signal is issued.

[0022] As a preferred technical solution, the improved Gaussian mixture model uses an adaptive learning rate to update parameters, that is, by setting a frame number threshold, when the frame number threshold is less than the frame number threshold, the background update rate is accelerated; when the frame number threshold is greater than the frame number threshold, the background update rate is slowed down;

[0023] The improved three-frame difference method performs an absolute difference operation on a grayscale image sequence of three consecutive frames in the target object's moving motion, obtains two binary images, performs a logical OR operation, and performs morphological processing. At the same time, the Sobel edge operator is introduced to supplement the edge contrast. The optimal number of frames participating in the edge contrast is adaptively obtained according to the target object's motion parameters, thereby reducing background misjudgment and obtaining the moving target pixel area.

[0024] As a preferred technical solution, the corresponding reconstruction of the three-dimensional point cloud scene in the two-dimensional image is achieved by:

[0025] Using the principle of laser ranging, raw point cloud data is collected through laser radar scanning. The point cloud data is loaded and drawn in the solid modeling tool to obtain a laser point cloud 3D model of the transmission line span and its channel range.

[0026] The target points are selected based on the transmission towers within the camera's field of view. The targets within the field of view are identified and matched one by one through target detection to obtain the target's three-dimensional feature points.

[0027] Based on the 3D point cloud and the camera captured image, the image pixels and 3D feature points are matched. The transformation principle between different coordinate systems is used to calculate the mapping relationship between the laser point cloud and the captured image, and the corresponding matching 3D scene in the captured image is obtained to achieve the fusion and correspondence of the two data.

[0028] As a preferred technical solution, the mapping relationship between the laser point cloud and the captured image is obtained as follows:

[0029] Define the laser point cloud 3D coordinate system and the 2D image coordinate system;

[0030] According to the principle of similar triangles, a relationship is established between a point in the 2D image coordinate system and the corresponding point in the 3D coordinate system of the laser point cloud;

[0031] Transform the two-dimensional image coordinate system of the point into the pixel coordinate system by scaling and translating the origin;

[0032] The three-dimensional coordinate system of the laser point cloud of the corresponding points is transformed into the world coordinate system through rotation and translation, and the three-dimensional point cloud is projected onto the two-dimensional image to become pixel points.

[0033] As a preferred technical solution, the generation of a three-dimensional point cloud alarm is specifically as follows:

[0034] Obtain camera captured images, perform target detection and extract target features;

[0035] Match target features in the 3D scene to reconstruct the target objects and their locations in the transmission line protection zone, including:

[0036] In a three-dimensional scene, each point has a unique spatial coordinate (x, y, z); in a two-dimensional image, each pixel has a unique plane coordinate (x, y) in the image;

[0037] Based on the mapping relationship between the laser point cloud and the captured image, the 3D spatial data model is reconstructed, and the target object and its position in the transmission line protection zone are reconstructed based on the pixel point set in the target feature.

[0038] Using the principle of similar triangles, the distance between multiple points in the three-dimensional space of the target object and the transmission line is calculated;

[0039] The shortest distance data between the transmission line and the target object is screened and compared with the safety distance of the corresponding voltage level; if it is less than the safety distance of the corresponding voltage level, a three-dimensional point cloud alarm is generated.

[0040] On the other hand, the present invention provides a transmission line external force damage identification and alarm system, characterized by comprising a target recognition module, an image recognition and alarm module, a video analysis and alarm module, a three-dimensional scene reconstruction module, a three-dimensional point cloud alarm module, and an alarm action execution module;

[0041] The target recognition module is used to classify target objects in the transmission line protection zone within the camera's field of view, and to extract and segment the camera-captured images based on target detection and convolutional neural networks to obtain target recognition results.

[0042] The image recognition alarm module uses the tower within the camera's field of view as the center line to demarcate the alarm area for the protection zone of the transmission line of different voltage levels and perform identification and selection. Combined with the target recognition results, it generates an image recognition alarm for the target object threatening the transmission line;

[0043] The video analysis and alarm module starts real-time video recording based on the image recognition alarm, obtains the target object's movement over a period of time, extracts video frames containing the target object, analyzes whether the target object is approaching the power transmission line in combination with target detection, and generates video analysis alarms using video background extraction and inter-frame difference;

[0044] The 3D scene reconstruction module obtains the laser point cloud 3D model of the transmission line span and its channel range, performs spatial coordinate transformation on the target within the camera field of view, and realizes the corresponding reconstruction of the 3D point cloud scene in the 2D image;

[0045] The 3D point cloud alarm module matches and reconstructs the target objects and their positions in the transmission line protection zone in the 3D scene, combines the principle of similar triangles to perform multi-directional distance measurement between the target objects and the transmission lines, selects the shortest distance, and compares it with the safe distance of the transmission line at the corresponding voltage level to generate a 3D point cloud alarm;

[0046] The alarm action execution module comprehensively analyzes the image recognition alarm, video analysis alarm and three-dimensional point cloud alarm, mutually verifies and detects the output of the transmission line external force damage alarm information, and if two or more of the three issue an alarm, the transmission line external force damage alarm action is immediately executed.

[0047] In another aspect, the present invention provides a computer-readable storage medium storing a program, wherein when the program is executed by a processor, the method for identifying and alarming external force damage to a power transmission line is implemented.

[0048] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0049] 1. The present invention uses a dual method of target detection and feature extraction to identify hidden dangers of external force damage in the protection zone of the transmission line. The recognition accuracy is high, which saves the time cost of the operation and maintenance personnel and improves the recognition efficiency.

[0050] 2. The present invention analyzes the approaching behavior of target objects by performing time differentiation and target detection on the target objects' movement over a period of time. This avoids false warnings in actual scenarios where some target objects are not within the safe distance of the power transmission line protection zone or are only briefly staying there without any dangerous approaching trend, thus ensuring the accuracy of the alarm.

[0051] 3. The present invention uses a laser point cloud 3D model to reconstruct a 3D scene, calculates and screens the shortest distance between the target object and the transmission line, generates a 3D point cloud alarm, combines image recognition alarms and video analysis alarms for comprehensive analysis, and executes external force damage alarm actions, thereby avoiding external force damage to the transmission line, reducing the missed alarm rate and false alarm rate of the alarm, and improving the alarm efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0053] Figure 1 This is a flow chart of a method for identifying and issuing an alarm for external force damage to a power transmission line according to an embodiment of the present invention;

[0054] Figure 2 This is a structural diagram of a system for identifying and warning of external force damage to a power transmission line according to an embodiment of the present invention;

[0055] Figure 3 Schematic diagram of the structure of a computer-readable storage medium in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0057] References to "embodiments" in this application mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described in this application may be combined with other embodiments.

[0058] like Figure 1 As shown, this embodiment provides a method for identifying and issuing an alarm for external force damage to a power transmission line, comprising the following steps:

[0059] S1. Classify the target objects in the transmission line protection zone within the camera's field of view, extract and segment the camera-captured images based on target detection and convolutional neural networks, and obtain target recognition results.

[0060] In this embodiment, the target detection adopts the YOLOv4 target detection network structure and uses the K-means algorithm to perform cluster analysis on the captured images;

[0061] The convolutional neural network is constructed based on Tensorflow and Faster-RCNN network. The hierarchical weighted cross entropy is used as the loss function to train the convolutional neural network, and the Focal Loss algorithm is used to optimize the loss function. The convolutional neural network is used for two learning and training sessions to extract target features and segment images to obtain target recognition results.

[0062] S2: Using the tower within the camera's field of view as the center line, demarcate alarm areas for the protection zones of transmission lines of different voltage levels and perform identification and selection. Combined with the target recognition results, generate image recognition alarms for target objects that threaten the transmission lines. Specifically:

[0063] Within the camera's field of view, the transmission line tower is used as the center line, and the safety distances of different voltage levels are used as parallel lines. The vertical area between the two parallel lines is defined as the alarm area.

[0064] Target detection is used to identify and select transmission line towers and protection zones of different voltage levels. Combined with the target recognition results, when the target object enters the alarm area, an image recognition alarm is generated for the target object that threatens the transmission line.

[0065] S3. Initiate real-time video recording based on the image recognition alarm, obtain the target object's movement over a period of time, extract the video frame image containing the target object, combine target detection to analyze whether the target object is approaching the power transmission line, and generate a video analysis alarm using video background extraction and inter-frame difference. Specifically:

[0066] The moving motion of the target object is verified and matched by improving the mixed Gaussian model to obtain the background image and extract the target;

[0067] The improved three-frame difference method is used to perform temporal difference between the image frame of the target object in motion and the background image to obtain the moving target pixel area;

[0068] The background image, the extracted target and the moving target pixel area are integrated, and the transmission line and tower are used as reference objects to analyze whether the target object is approaching the transmission line. If so, a video analysis alarm is generated and a warning signal is issued.

[0069] The traditional Gaussian mixture modeling model is updated with a fixed learning rate. The specific steps are:

[0070] Each pixel in the background image is composed of K Gaussian distributions, so K mixed Gaussian distribution probability density functions are defined; the weight coefficient, mean vector and covariance of each mixed Gaussian distribution probability density function in the improved mixed Gaussian model are initialized; the pixel values ​​of the image frames at different times during the target object's movement are matched with the corresponding mixed Gaussian distribution probability density function, and if the match is qualified, the mixed Gaussian distribution probability density function is defined to match the pixel value; for the matched mixed Gaussian distribution probability density function, its mean vector and covariance matrix are parameter updated at a fixed learning rate; for the unmatched mixed Gaussian distribution probability density function, its mean vector and covariance matrix remain unchanged; determine which mixed Gaussian distribution probability density functions in the mixed Gaussian model are matched by the background pixels, and select the first N mixed Gaussian distribution probability density functions as the background pixel model based on the mixed Gaussian distribution probability density function sequence; re-check the matching relationship between the pixel values ​​of the image frames at different times during the target object's movement and the background pixel model, and if they match, they are background points; after the re-check is completed, the background image is accurately extracted. However, the traditional Gaussian mixture model is prone to target loss and misjudgment.

[0071] The improved Gaussian mixture model in this embodiment uses an adaptive learning rate for parameter update, which solves the problems of target loss and misjudgment. By setting a frame number threshold, when the frame number threshold is less than the frame number threshold, the background update rate is accelerated, and when it is greater than the frame number threshold, the background update rate is slowed down, thereby achieving accurate background image extraction;

[0072] Temporal differencing involves subtracting images from adjacent time points in a target scene, thereby capturing the temporal transformation of the target scene. The improved three-frame differencing method employed in this embodiment performs an absolute difference operation on a sequence of three consecutive grayscale images of the target object during its motion. The resulting two binary images are then subjected to a logical OR operation and morphological processing. A Sobel edge operator is then introduced for edge comparison. The optimal number of frames participating in edge comparison is adaptively obtained based on the target motion parameters, reducing the misjudgment rate of background points and eliminating edge blurring caused by changes in motion position, ensuring that the edges of the moving target are continuous and complete. Edges are then supplemented on the three-frame difference results, and noise interference is eliminated through morphological filtering to obtain the pixel region of the moving target.

[0073] S4. Obtain a three-dimensional laser point cloud model of the transmission line span and its channel range, perform spatial coordinate transformation on the target within the camera field of view, and achieve corresponding reconstruction of the three-dimensional point cloud scene in the two-dimensional image. Specifically:

[0074] Using the principle of laser ranging, raw point cloud data is collected through laser radar scanning. The point cloud data is loaded and drawn in the solid modeling tool to obtain a laser point cloud 3D model of the transmission line span and its channel range.

[0075] The target points are selected based on the transmission towers within the camera's field of view. The targets within the field of view are identified and matched one by one through target detection to obtain the target's three-dimensional feature points.

[0076] Based on the 3D point cloud and the camera captured image, the image pixels and 3D feature points are matched. The transformation principle between different coordinate systems is used to calculate the mapping relationship between the laser point cloud and the captured image, and the corresponding matching 3D scene in the captured image is obtained. The fusion correspondence of the two data is realized to achieve the goal of 3D scene reconstruction.

[0077] In this embodiment, the mapping relationship between the laser point cloud and the captured image is:

[0078] Define the point cloud 3D coordinate system and the 2D image coordinate system; establish a relationship between a point in the 2D image coordinate system and the corresponding point in the 3D coordinate system of the laser point cloud based on the principle of similar triangles; transform the 2D image coordinate system where the point is located into the pixel coordinate system through scaling and origin translation; uniformly transform the 3D coordinate system of the laser point cloud into the world coordinate system through rotation and translation; complete the projection of the 3D point cloud onto the 2D image and convert it into pixel points.

[0079] S5. Based on the target objects and their locations in the transmission line protection zone that are matched and reconstructed in the 3D scene, the multi-directional distance measurement between the target objects and the transmission lines is performed using the principle of similar triangles. The shortest distance is selected and compared with the safe distance of the transmission line at the corresponding voltage level to generate a 3D point cloud alarm. Specifically:

[0080] First, obtain the camera captured image, perform target detection and extract target features;

[0081] Then, the target features are matched in the 3D scene to reconstruct the target objects and their locations in the transmission line protection zone. The principle is as follows:

[0082] In a 3D scene, each point has unique spatial coordinates (x, y, z); in a 2D image, each pixel has unique plane coordinates (x, y). Based on the mapping relationship between the laser point cloud and the captured image, a 3D spatial data model is reconstructed. The target objects and their locations in the transmission line protection zone are reconstructed based on the target pixel point set.

[0083] Then, the principle of similar triangles is used to calculate the distances between multiple points in the three-dimensional space of the target object and the transmission line;

[0084] Finally, the shortest distance data between the transmission line and the target object is screened and compared with the safety distance of the corresponding voltage level; if it is less than the safety distance of the corresponding voltage level, a three-dimensional point cloud alarm is generated.

[0085] In this embodiment, the safety distance corresponding to the voltage level refers to the safety distance between the live conductor of the transmission line and the target object. For example, the safety distance between the live conductor of a 110kV transmission line and the target object is 5m; the safety distance between the live conductor of a 220kV transmission line and the target object is 6m; and the safety distance between the live conductor of a 500kV transmission line and the target object is 8.5m.

[0086] S6. Comprehensively analyze the image recognition alarm, video analysis alarm and three-dimensional point cloud alarm, mutually verify the detection and output of the transmission line external force damage alarm information. If two or more of the three issue an alarm, the transmission line external force damage alarm action will be executed immediately. Otherwise, the transmission line protection zone will continue to monitor the external force damage.

[0087] It should be noted that, for the sake of convenience, the aforementioned method embodiments are all expressed as a series of action combinations, but those skilled in the art should know that the present invention is not limited to the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously.

[0088] Based on the same concept as the method for identifying and alarming external force damage to a transmission line in the above-mentioned embodiment, the present invention also provides a system for identifying and alarming external force damage to a transmission line, which can be used to implement the above-mentioned method for identifying and alarming external force damage to a transmission line. For ease of explanation, the structural diagram of an embodiment of a system for identifying and alarming external force damage to a transmission line only shows the parts relevant to the embodiment of the present invention. Those skilled in the art will understand that the illustrated structure does not constitute a limitation of the device, and the device may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0089] like Figure 2 As shown, another embodiment of the present invention provides a recognition and alarm system for external force damage to a power transmission line, including a target recognition module, an image recognition and alarm module, a video analysis and alarm module, a three-dimensional scene reconstruction module, a three-dimensional point cloud alarm module, and an alarm action execution module;

[0090] The target recognition module is used to classify target objects in the transmission line protection zone within the camera's field of view. It extracts and segments features of the camera-captured images based on target detection and convolutional neural networks to obtain target recognition results.

[0091] The image recognition alarm module uses the tower within the camera's field of view as the center line to demarcate alarm areas for transmission line protection zones of different voltage levels and perform identification and selection. Combined with the target recognition results, it generates image recognition alarms for target objects that threaten transmission lines.

[0092] The video analysis and alarm module initiates real-time video recording based on image recognition alarms, captures the target object's movement over a period of time, extracts video frames containing the target object, analyzes whether the target object is approaching the power transmission line through target detection, and generates video analysis alarms using video background extraction and inter-frame difference.

[0093] The 3D scene reconstruction module obtains the 3D laser point cloud model of the transmission line span and its channel range, and performs spatial coordinate transformation corresponding to the target within the camera's field of view to achieve corresponding reconstruction of the 3D point cloud scene in the 2D image;

[0094] The 3D point cloud alarm module matches and reconstructs the target objects and their locations in the transmission line protection zone in the 3D scene. It then uses the principle of similar triangles to perform multi-directional distance measurements between the target objects and the transmission lines. It then selects the shortest distance and compares it with the safe distance of the transmission line at the corresponding voltage level to generate a 3D point cloud alarm.

[0095] The alarm action execution module comprehensively analyzes the image recognition alarm, video analysis alarm and three-dimensional point cloud alarm, mutually verifies and detects the output of the transmission line external force damage alarm information. If two or more of the three issue an alarm, the transmission line external force damage alarm action will be executed immediately.

[0096] It should be noted that a system for identifying and alarming external force damage to a transmission line of the present invention corresponds one-to-one to a method for identifying and alarming external force damage to a transmission line of the present invention. The technical features and beneficial effects described in the embodiment of the above-mentioned method for identifying and alarming external force damage to a transmission line are applicable to the embodiment of a system for identifying and alarming external force damage to a transmission line. For specific contents, please refer to the description in the embodiment of the method of the present invention. No further details will be given here. This is hereby declared.

[0097] In addition, in the above-mentioned implementation of the transmission line external force damage identification and alarm system, the logical division of each program module is only an example. In actual application, the above-mentioned functions can be assigned to different program modules as needed, for example, for the configuration requirements of the corresponding hardware or the convenience of software implementation. That is, the internal structure of the transmission line external force damage identification and alarm system is divided into different program modules to complete all or part of the functions described above.

[0098] like Figure 3As shown, in one embodiment, a computer-readable storage medium is provided, which stores a program in a memory. When the program is executed by a processor, the method for identifying and alarming external force damage to a power transmission line is implemented, specifically:

[0099] Classify the target objects in the transmission line protection zone within the camera's field of view, extract and segment the camera-captured images based on target detection and convolutional neural networks, and obtain target recognition results.

[0100] Using the tower within the camera's field of view as the center line, the system demarcates alarm areas for transmission line protection zones of different voltage levels and performs identification and selection. Combined with the target recognition results, it generates image recognition alarms for target objects threatening transmission lines.

[0101] Initiate real-time video recording based on image recognition alarms, capture the target object's movement over a period of time, extract video frames containing the target object, combine target detection to analyze whether the target object is approaching the power transmission line, and generate video analysis alarms using video background extraction and inter-frame difference.

[0102] Obtain a 3D laser point cloud model of the transmission line span and its channel range, perform spatial coordinate transformation on the target within the camera's field of view, and achieve corresponding reconstruction of the 3D point cloud scene in the 2D image;

[0103] Based on the target objects and their locations in the transmission line protection zone that are matched and reconstructed in the 3D scene, the multi-directional distance measurement between the target objects and the transmission lines is performed using the principle of similar triangles. The shortest distance is selected and compared with the safe distance of the transmission line at the corresponding voltage level to generate a 3D point cloud alarm.

[0104] Comprehensively analyze image recognition alarms, video analysis alarms and three-dimensional point cloud alarms, mutually verify and detect the output of transmission line external force damage alarm information. If two or more of the three issue an alarm, the transmission line external force damage alarm action will be immediately executed.

[0105] Those skilled in the art will appreciate that all or part of the processes in the above-described embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-described methods. In particular, any reference to memory, storage, database, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory.

[0106] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0107] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be considered as equivalent replacement methods and are included in the scope of protection of the present invention.

Claims

1. A method for identifying and warning of external force damage to a power transmission line, characterized in that: The steps include: Classify the target objects in the transmission line protection zone within the camera's field of view, extract and segment the camera-captured images based on target detection and convolutional neural networks, and obtain target recognition results. The target detection adopts the YOLOv4 target detection network structure and uses the K-means algorithm to perform cluster analysis on the captured images. The convolutional neural network is constructed based on TensorFlow and the Faster-RCNN network. The convolutional neural network is trained using hierarchical weighted cross entropy as the loss function, and the Focal Loss algorithm is used to optimize the loss function. The convolutional neural network is used for two learning and training sessions to extract target features and segment the image to obtain target recognition results. Using the tower within the camera's field of view as the center line, the system demarcates alarm areas for transmission line protection zones of different voltage levels and performs identification and selection. Combined with the target recognition results, it generates image recognition alarms for target objects threatening transmission lines. The generating of the image recognition alarm is specifically as follows: Within the camera's field of view, the transmission line tower is used as the center line, and the safety distances of different voltage levels are used as parallel lines. The vertical area between the two parallel lines is defined as the alarm area. Target detection is used to identify and select transmission line towers and protected areas of different voltage levels. When a target object enters the alarm area, an image recognition alarm is generated for the target object threatening the transmission line. Initiate real-time video recording based on image recognition alarms, capture the target object's movement over a period of time, extract video frames containing the target object, combine target detection to analyze whether the target object is approaching the power transmission line, and generate video analysis alarms using video background extraction and inter-frame difference. Obtain a 3D laser point cloud model of the transmission line span and its channel range, perform spatial coordinate transformation on the target within the camera's field of view, and achieve corresponding reconstruction of the 3D point cloud scene in the 2D image; Based on the target objects and their locations in the transmission line protection zone that are matched and reconstructed in the 3D scene, the multi-directional distance measurement between the target objects and the transmission lines is performed using the principle of similar triangles. The shortest distance is selected and compared with the safe distance of the transmission line at the corresponding voltage level to generate a 3D point cloud alarm. Comprehensively analyze image recognition alarms, video analysis alarms and three-dimensional point cloud alarms, mutually verify and detect the output of transmission line external force damage alarm information. If two or more of the three issue an alarm, the transmission line external force damage alarm action will be immediately executed.

2. The method for identifying and warning of external force damage to a power transmission line according to claim 1, characterized in that: The generating of the video analysis alarm is specifically as follows: The moving motion of the target object is verified and matched by improving the mixed Gaussian model to obtain the background image and extract the target; The improved three-frame difference method is used to perform temporal difference between the image frame of the target object in motion and the background image to obtain the moving target pixel area; The background image, the extracted target and the moving target pixel area are integrated, and the transmission line and tower are used as reference objects to analyze whether the target object is approaching the transmission line. If so, a video analysis alarm is generated and a warning signal is issued.

3. The method for identifying and warning of external force damage to a power transmission line according to claim 2, characterized in that: The improved Gaussian mixture model uses an adaptive learning rate to update parameters, that is, by setting a frame number threshold, when the frame number threshold is less than the frame number threshold, the background update rate is accelerated; when the frame number threshold is greater than the frame number threshold, the background update rate is slowed down; The improved three-frame difference method performs an absolute difference operation on a grayscale image sequence of three consecutive frames in the target object's moving motion, obtains two binary images, performs a logical OR operation, and performs morphological processing. At the same time, the Sobel edge operator is introduced to supplement the edge contrast. The optimal number of frames participating in the edge contrast is adaptively obtained according to the target object's motion parameters, thereby reducing background misjudgment and obtaining the moving target pixel area.

4. The method for identifying and warning of external force damage to a power transmission line according to claim 2, characterized in that: The corresponding reconstruction of the three-dimensional point cloud scene in the two-dimensional image is specifically achieved as follows: Using the principle of laser ranging, raw point cloud data is collected through laser radar scanning. The point cloud data is loaded and drawn in the solid modeling tool to obtain a laser point cloud 3D model of the transmission line span and its channel range. The target points are selected based on the transmission towers within the camera's field of view. The targets within the field of view are identified and matched one by one through target detection to obtain the target's three-dimensional feature points. Based on the 3D point cloud and the camera captured image, the image pixels and 3D feature points are matched. The transformation principle between different coordinate systems is used to calculate the mapping relationship between the laser point cloud and the captured image, and the corresponding matching 3D scene in the captured image is obtained to achieve the fusion and correspondence of the two data.

5. The method for identifying and warning of external force damage to a power transmission line according to claim 4, characterized in that: The mapping relationship between the laser point cloud and the captured image is obtained as follows: Define the laser point cloud 3D coordinate system and the 2D image coordinate system; According to the principle of similar triangles, a relationship is established between a point in the 2D image coordinate system and the corresponding point in the 3D coordinate system of the laser point cloud; Transform the two-dimensional image coordinate system of the point into the pixel coordinate system by scaling and translating the origin; The three-dimensional coordinate system of the laser point cloud of the corresponding points is transformed into the world coordinate system through rotation and translation, and the three-dimensional point cloud is projected onto the two-dimensional image to become pixel points.

6. The method for identifying and warning of external force damage to a power transmission line according to claim 5, characterized in that: The generation of a three-dimensional point cloud alarm is specifically as follows: Obtain camera captured images, perform target detection and extract target features; Match target features in the 3D scene to reconstruct the target objects and their locations in the transmission line protection zone, including: In a three-dimensional scene, each point has a unique spatial coordinate (x, y, z); in a two-dimensional image, each pixel has a unique plane coordinate (x, y) in the image; Based on the mapping relationship between the laser point cloud and the captured image, the 3D spatial data model is reconstructed, and the target object and its position in the transmission line protection zone are reconstructed based on the pixel point set in the target feature. Using the principle of similar triangles, the distance between multiple points in the three-dimensional space of the target object and the transmission line is calculated; The shortest distance data between the transmission line and the target object is screened and compared with the safety distance of the corresponding voltage level; if it is less than the safety distance of the corresponding voltage level, a three-dimensional point cloud alarm is generated.

7. A system for identifying and warning of external force damage to power transmission lines, characterized in that: It includes target recognition module, image recognition alarm module, video analysis alarm module, 3D scene reconstruction module, 3D point cloud alarm module and alarm action execution module; The target recognition module is used to classify target objects in the transmission line protection zone within the camera's field of view, and to extract and segment the camera-captured images based on target detection and convolutional neural networks to obtain target recognition results. The target detection adopts the YOLOv4 target detection network structure and uses the K-means algorithm to perform cluster analysis on the captured images. The convolutional neural network is constructed based on TensorFlow and the Faster-RCNN network. The convolutional neural network is trained using hierarchical weighted cross entropy as the loss function, and the Focal Loss algorithm is used to optimize the loss function. The convolutional neural network is used for two learning and training sessions to extract target features and segment the image to obtain target recognition results. The image recognition alarm module uses the tower within the camera's field of view as the center line to demarcate the alarm area for the protection zone of the transmission line of different voltage levels and perform identification and selection. Combined with the target recognition results, it generates an image recognition alarm for the target object threatening the transmission line; The generating of the image recognition alarm is specifically as follows: Within the camera's field of view, the transmission line tower is used as the center line, and the safety distances of different voltage levels are used as parallel lines. The vertical area between the two parallel lines is defined as the alarm area. Target detection is used to identify and select transmission line towers and protected areas of different voltage levels. When a target object enters the alarm area, an image recognition alarm is generated for the target object threatening the transmission line. The video analysis and alarm module starts real-time video recording based on the image recognition alarm, obtains the target object's movement over a period of time, extracts video frames containing the target object, analyzes whether the target object is approaching the power transmission line in combination with target detection, and generates video analysis alarms using video background extraction and inter-frame difference; The 3D scene reconstruction module obtains the laser point cloud 3D model of the transmission line span and its channel range, performs spatial coordinate transformation on the target within the camera field of view, and realizes the corresponding reconstruction of the 3D point cloud scene in the 2D image; The 3D point cloud alarm module matches and reconstructs the target objects and their positions in the transmission line protection zone in the 3D scene, combines the principle of similar triangles to perform multi-directional distance measurement between the target objects and the transmission lines, selects the shortest distance, and compares it with the safe distance of the transmission line at the corresponding voltage level to generate a 3D point cloud alarm; The alarm action execution module comprehensively analyzes the image recognition alarm, video analysis alarm and three-dimensional point cloud alarm, mutually verifies and detects the output of the transmission line external force damage alarm information, and if two or more of the three issue an alarm, the transmission line external force damage alarm action is immediately executed.

8. A computer-readable storage medium storing a program, characterized in that: When the program is executed by a processor, the method for identifying and warning of external force damage to a power transmission line as described in any one of claims 1 to 6 is implemented.

Images