Anti-electric shock warning method and system based on monitoring video network data

Abstract

The invention discloses an electric shock prevention warning method and system based on monitoring video network data, and relates to the technical field of monitoring video data analysis, and the method comprises the steps: dividing a basic dangerous region of target power equipment according to traffic flow information and electric shock historical accident information near the target power equipment; and the size of a basic dangerous area is determined by considering the previous traffic flow and accident occurrence condition of the target power equipment, so that a reliable basis is provided for subsequent warning of the power equipment. And establishing a template of the target power equipment according to the inherent image features. The method comprises the following steps: identifying and predicting the moving information of a moving object around target power equipment, predicting the condition that the moving object enters a basic dangerous area, and adjusting the basic dangerous area of the target power equipment according to the moving information of the moving object and the current state of the target power equipment, thereby adaptively early warning and warning. Electric shock prevention dynamic alert and automatic electric shock prevention alert are achieved, and safety of pedestrians and vehicles and normal operation of power equipment are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of monitoring video data analysis, and particularly to an anti-electric shock warning method and system based on monitoring video network data. Background Art

[0002] With the increasing complexity and scale expansion of the power network, the safe operation of power facilities faces many challenges, especially the risk of electric shock accidents cannot be ignored. Although traditional power monitoring systems can achieve data collection and remote control, they have limitations in anti-electric shock early warning. Therefore, by combining modern video monitoring technology and network data transmission, and through real-time analysis of the attributes of moving objects (such as personnel, vehicles, etc.) and dynamic information (such as moving speed, trajectory) in the monitoring video, potential electric shock risks can be accurately identified.

[0003] In the prior art, the anti-electric shock warning of power equipment often sets warning signs to prompt pedestrians or vehicles of danger, and relies on people's self-awareness to achieve warning. Or, a relatively large fixed warning range is set, and when a moving object is detected within the warning range, a warning is given. The first method can only rely on people's self-awareness to prevent electric shock, resulting in poor automaticity of anti-electric shock warning and low warning effect. The second method cannot dynamically adjust the warning range according to the actual situation, and does not meet the complex anti-electric shock requirements of power equipment in public areas, resulting in poor warning adaptability and low warning accuracy.

[0004] Therefore, how to improve the automaticity, adaptability and accuracy of anti-electric shock warning is a technical problem to be solved at present. Summary of the Invention

[0005] The purpose of the present invention is to solve the problems of poor automaticity, poor adaptability and low accuracy of anti-electric shock warning in the prior art, and to propose an anti-electric shock warning method based on monitoring video network data, which includes, Determine the location of the target power equipment, and collect traffic flow information and electric shock historical accident information near the target power equipment, so as to divide the basic dangerous area of the target power equipment; Screen out the inherent image features of the target power equipment in the monitoring video, establish a template of the target power equipment by virtue of the inherent image features, and identify the target power equipment and its current state in the monitoring video through the template of the target power equipment; Obtain the real-time monitoring video around the target power equipment, identify the attributes and dynamic information of the moving object in the real-time monitoring video, and predict the moving information of the moving object based on the attributes and dynamic information of the moving object; Adjust the basic danger area of the target power equipment according to the movement information of the moving object and the current state of the target power equipment to obtain the danger area of the target power equipment. When the moving object is in the danger area, a warning is issued around the target power equipment to achieve dynamic anti-electric shock warning.

[0006] In some embodiments of the present application, traffic flow information and electric shock historical accident information near the target power equipment are collected, and based on this, the basic danger area of the target power equipment is divided, including: Collect traffic flow information and electric shock historical accident information within a preset range near the target power equipment, respectively construct a traffic flow distribution map and an electric shock historical accident distribution map, perform position alignment processing on the traffic flow distribution map and the electric shock historical accident distribution map, and overlay the traffic flow distribution map and the electric shock historical accident distribution map to generate a traffic flow - electric shock accident distribution map. The corresponding positions on the traffic flow - electric shock accident distribution map are marked with traffic flow characteristics and electric shock accident event parameters. Determine the distribution degree of electric shock accidents through the positions in the electric shock accident event parameters on the traffic flow - electric shock accident distribution map, perform grid processing on the preset range near the target power equipment according to the distribution degree of electric shock accidents, and divide the area near the target power equipment into two categories: the area where electric shock has occurred and the area where electric shock has not occurred. Calculate the distances from the area where electric shock has occurred and the area where electric shock has not occurred to the target power equipment respectively. Respectively determine the regional risk levels of the area where electric shock has occurred and the area where electric shock has not occurred, and combine the regional risk levels of the area where electric shock has occurred and the area where electric shock has not occurred to determine the basic danger area of the target power equipment.

[0007] In some embodiments of the present application, respectively determine the regional risk levels of the area where electric shock has occurred and the area where electric shock has not occurred, including: For the area where electric shock has occurred, confirm the average value of the traffic flow characteristics of the grid and adjacent grids in the area where electric shock has occurred, calculate the ratio of the area of the grid in the area where electric shock has occurred to the area within the preset range near the target power equipment, and determine the regional risk level of the area where electric shock has occurred based on the average value of the traffic flow characteristics, the distance from the area where electric shock has occurred to the target power equipment, the electric shock accident event parameters in the area where electric shock has occurred, and the ratio of the area of the grid in the area where electric shock has occurred to the area within the preset range near the target power equipment. For the area where electric shock has not occurred, calculate the distance between the area where electric shock has not occurred and the nearest area where electric shock has occurred, denoted as the nearest distance, calculate the ratio of the area of the grid in the area where electric shock has not occurred to the area within the preset range near the target power equipment, and determine the regional risk level of the area where electric shock has not occurred based on the traffic flow characteristics of the area where electric shock has not occurred, the nearest distance, and the distance from the area where electric shock has not occurred to the target power equipment.

[0008] In some embodiments of the present application, the inherent image features of the target power equipment on the monitoring video are screened out, and a template of the target power equipment is established based on the inherent image features, including: Extract the image features on the monitoring video according to the structure of the target power equipment and the identifiers of the target power equipment, and perform feature fusion on different types of image features to generate a fused feature vector. Construct a classification model according to the different operating states of the target power equipment, and establish a template of the target power equipment based on the fused feature vector and the classification model.

[0009] In some embodiments of the present application, before identifying the attributes and dynamic information of the moving object in the real-time monitoring video, the method further includes: Search for multiple fixed references around the target power equipment, determine the actual position coordinates of the fixed references and the pixel position coordinates of the fixed references in the monitoring video, and determine a coordinate conversion scale ratio, and convert the pixel position coordinates in the monitoring video into actual position coordinates through the coordinate conversion scale ratio.

[0010] In some embodiments of the present application, identifying the attributes and dynamic information of the moving object in the real-time monitoring video includes: Identify the attributes and contour size of the moving object in the monitoring video through an identification model, and the attributes of the moving object include the category and contour size of the moving object. The dynamic information of the moving object includes the moving speed and moving trajectory, and the moving speed and moving trajectory of the moving object in the monitoring video are identified through the optical flow method and Kalman filtering.

[0011] In some embodiments of the present application, predicting the movement information of the moving object based on the attributes and dynamic information of the moving object includes: Input the attributes and dynamic information of the moving object within a current period of time into the moving object prediction model, and the moving object prediction model outputs the movement information of the moving object within a future period of time.

[0012] In some embodiments of the present application, adjusting the basic danger area of the target power equipment according to the movement information of the moving object and the current state of the target power equipment to obtain the danger area of the target power equipment includes: Calculate the intersection degree between the moving object and the basic danger area of the target power equipment according to the movement information of the moving object, and adjust the basic danger area of the target power equipment according to the intersection degree and the current state of the target power equipment to obtain the danger area of the target power equipment.

[0013] Correspondingly, the present application also provides an anti-electric shock warning system based on monitoring video network data, including: The first module is used to determine the location of the target power equipment, collect traffic flow information and electric shock historical accident information near the target power equipment, and thereby divide the basic dangerous area of the target power equipment; The second module is used to screen out the inherent image features of the target power equipment in the surveillance video, establish a template of the target power equipment by virtue of the inherent image features, and identify the target power equipment and its current state in the surveillance video through the template of the target power equipment; The third module is used to obtain the real-time surveillance video around the target power equipment, identify the attributes and dynamic information of the moving objects in the real-time surveillance video, and predict the moving information of the moving objects based on the attributes and dynamic information of the moving objects; The fourth module is used to adjust the basic dangerous area of the target power equipment according to the moving information of the moving objects and the current state of the target power equipment to obtain the dangerous area of the target power equipment. When the moving object is in the dangerous area, a warning is issued to the periphery of the target power equipment, thereby realizing dynamic anti-electric shock warning.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The basic dangerous area of the target power equipment is divided according to the traffic flow information and electric shock historical accident information near the target power equipment. Considering the past traffic flow and accident occurrence situation of the target power equipment, the size of a basic dangerous area is determined, providing a reliable basis for the subsequent warning of the power equipment. A template of the target power equipment is established according to the inherent image features, and thereby the current state of the target power equipment is identified in the video surveillance, facilitating the subsequent adjustment of the warning range.

[0015] 2. The moving information of the moving objects around the target power equipment is identified and predicted to predict the situation of the moving objects entering the basic dangerous area. The basic dangerous area of the target power equipment is adjusted according to the moving information of the moving objects and the current state of the target power equipment, thereby realizing adaptive early warning and warning, achieving dynamic anti-electric shock warning and automatic anti-electric shock warning, improving the accuracy and adaptability of the anti-electric shock warning of the power equipment, reducing the probability of warning errors, and ensuring the safety of pedestrians and vehicles and the normal operation of the power equipment. Description of the Drawings

[0016] Figure 1 It is a schematic flow chart of the anti-electric shock warning method based on the surveillance video network data proposed by the present invention; Figure 2 It is a schematic structural diagram of the anti-electric shock warning system based on the surveillance video network data proposed by the present invention. Detailed Embodiments

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0018] Referring to Figure 1 , an anti-electric shock warning method based on monitoring video network data includes the following steps: Step S101, determine the location of the target power equipment, and collect traffic flow information and electric shock historical accident information near the target power equipment, so as to divide the basic dangerous area of the target power equipment.

[0019] In this embodiment, in the prior art, a relatively large fixed warning range is set. When a moving object is sensed within the warning range, sound and light warnings are given. This will cause warnings to be issued even for a little "stir" within the warning range. After a long time, the warning effect is very poor and the warning error rate is high. The fixed warning range is not suitable for preventing electric shock of power equipment in public areas, and the warning adaptability and accuracy are very poor.

[0020] In this embodiment, the longitude and latitude coordinates of the target power equipment are obtained through the GIS system, and a three-dimensional space model is established in combination with the equipment type (such as substation, transmission line). An initial acquisition area with a preset radius is delimited with the equipment as the center, and the historical traffic flow data (including traffic volume, pedestrian density, congestion index, etc.) of this area is obtained by calling the API of the traffic department. At the same time, the electric shock accident records (time, location, frequency, severity, etc.) of the power department are integrated, and a basic dangerous area is set considering the traffic flow and electric shock accidents.

[0021] In some embodiments of the present application, traffic flow information and electric shock historical accident information near the target power equipment are collected, and the basic dangerous area of the target power equipment is divided accordingly, including, Collect traffic flow information and electric shock historical accident information within a preset range near the target power equipment, respectively construct a traffic flow distribution map and an electric shock historical accident distribution map, perform position alignment processing on the traffic flow distribution map and the electric shock historical accident distribution map, and overlay the traffic flow distribution map and the electric shock historical accident distribution map to generate a traffic flow-electric shock accident distribution map. The corresponding positions on the traffic flow-electric shock accident distribution map are marked with traffic flow characteristics and electric shock accident event parameters; Determine the distribution degree of electric shock accidents through the position in the electric shock accident event parameters on the traffic flow-electric shock accident distribution map, perform grid processing on the preset range near the target power equipment according to the distribution degree of electric shock accidents, and divide the area near the target power equipment into two categories: the area where electric shock has occurred and the area where electric shock has not occurred; Calculate the distances between the area where electric shock has occurred and the area where electric shock has not occurred and the target power equipment respectively. Respectively determine the regional risk levels of the areas where electric shock has occurred and the areas where electric shock has not occurred, and determine the basic dangerous area of the target power equipment by combining the regional risk levels of the areas where electric shock has occurred and the areas where electric shock has not occurred.

[0022] In this embodiment, the distribution of electric shock accidents near the power equipment is statistically analyzed. The more dispersed the distribution is, the finer the grid degree is. It is split into multiple grids. According to whether an electric shock accident has occurred, it is divided into two categories: the area where electric shock has occurred and the area where electric shock has not occurred. Both categories of areas are composed of multiple grids. Determine the distance from each to the power equipment, and evaluate the regional risk levels of the two categories of areas, so as to set the size of the basic dangerous area of the power equipment. Different regional risk levels of the two categories of areas correspond to different sizes of the basic dangerous area.

[0023] In some embodiments of the present application, the regional risk levels of the areas where electric shock has occurred and the areas where electric shock has not occurred are respectively determined, including For the area where electric shock has occurred, confirm the average value of the traffic flow characteristics of the grids and adjacent grids in the area where electric shock has occurred, calculate the ratio of the area of the grids in the area where electric shock has occurred to the area within the preset range near the target power equipment, and determine the regional risk level of the area where electric shock has occurred based on the average value of the traffic flow characteristics, the distance from the area where electric shock has occurred to the target power equipment, the electric shock accident event parameters in the area where electric shock has occurred, and the ratio of the area of the grids in the area where electric shock has occurred to the area within the preset range near the target power equipment; For the area where electric shock has not occurred, calculate the distance between the area where electric shock has not occurred and the nearest area where electric shock has occurred, denoted as the nearest distance, calculate the ratio of the area of the grids in the area where electric shock has not occurred to the area within the preset range near the target power equipment, and determine the regional risk level of the area where electric shock has not occurred based on the traffic flow characteristics of the area where electric shock has not occurred, the nearest distance, and the distance from the area where electric shock has not occurred to the target power equipment.

[0024] In this embodiment, for the area where electric shock has occurred, the electric shock may occur at a point. In order to capture the traffic flow situation at that point, the average value of the traffic flow characteristics is calculated by combining the adjacent grids in the area where electric shock has occurred. The calculation formula for the regional risk level (Risk level of the area where an electric shock has occurred) of the area where electric shock has occurred is as follows: ; Wherein, is the regional risk level of the area where electric shock has occurred, 、 、 They are respectively the risk conversion coefficients of the traffic flow characteristics, the distance between the area where an electric shock has occurred and the target power equipment, and the ratio of the grid area under the area where an electric shock has occurred to the area within a preset range near the target power equipment. is the quantity of traffic flow characteristics, is the combined weight of the th traffic flow characteristic, is the magnitude of the th traffic flow characteristic, is the distance between the area where an electric shock has occurred and the target power equipment, is the ratio of the grid area under the area where an electric shock has occurred to the area within a preset range near the target power equipment, is the severity of the electric shock accident obtained from the electric shock accident parameters, , are two constants respectively, and [] is the rounding symbol. represents the correction of the severity of the electric shock accident to the sum of several aspects of risks. The two constants are respectively to balance the magnitude of the correction function and the magnitude of the regional risk level. Based on the average value of traffic flow characteristics, the distance between the area where an electric shock has occurred and the target power equipment, the electric shock accident parameters under the area where an electric shock has occurred, and the ratio of the grid area under the area where an electric shock has occurred to the area within a preset range near the target power equipment, these factors will affect the risk level of the region, and these contents are considered to evaluate the risk level.

[0025] In this embodiment, for the area where an electric shock has not occurred, based on the traffic flow characteristics, the closest distance, and the distance between the area where an electric shock has not occurred and the target power equipment of the area where an electric shock has not occurred, the regional risk level of the area where an electric shock has not occurred is determined. These factors will affect the risk level assessment of this type of area. Since no electric shock accident has occurred in this area, the distance closest to the area where an electric shock has occurred in this area is statistically counted to describe the influence of the area where an electric shock has occurred on it. From the attenuation principle, the closer to the power equipment or the closer to the area where an electric shock has occurred, the greater the risk. The calculation formula for the regional risk level (Risklevel of the area where an electric shock has not occurred) of the area where an electric shock has not occurred is as follows: ; Among them, is the regional risk level of the area where an electric shock has not occurred, , are respectively the risk conversion coefficients of the traffic flow characteristics and the distance between the area where an electric shock has not occurred and the target power equipment, is the quantity of traffic flow characteristics of the area where an electric shock has not occurred, is the The combined weight of traffic flow characteristics is the size of the traffic flow characteristic 、 are respectively the distance and the closest distance from the area where no electric shock occurs to the target power equipment and are preset constants with the same function as above represents the correction of the sum of risks by the closest distance

[0026] Step S102: Screen out the inherent image features of the target power equipment in the monitoring video, establish a template of the target power equipment based on the inherent image features, and identify the target power equipment and its current state in the monitoring video through the template of the target power equipment

[0027] In this embodiment, the image features of the power equipment in the monitoring video are determined to identify the power equipment and its state (such as whether it is abnormal) in the monitoring video

[0028] In some embodiments of the present application, screening out the inherent image features of the target power equipment in the monitoring video and establishing a template of the target power equipment based on the inherent image features includes extracting the image features on the monitoring video according to the structure of the target power equipment and the identifiers of the target power equipment, and performing feature fusion on different types of image features to generate a fused feature vector constructing a classification model according to the different operating states of the target power equipment, and establishing a template of the target power equipment based on the fused feature vector and the classification model

[0029] In this embodiment, the image features include device contour features, insulator features, warning sign features, transmission line features, etc. The Hough transform is used to detect rectangular structures. For devices such as transformers and switchgear in a substation, their appearances usually have rectangular features. After edge detection of the surveillance video frames through the Hough transform function in OpenCV, the Hough transform is applied to detect the rectangular contours and extract the contour features of the devices. The HOG (Histogram of Oriented Gradient) feature is used to extract the texture and shape information of the insulators. The HOG feature can effectively describe the edges and gradient distributions of objects in the image. Using the color threshold segmentation method, the surveillance video frames are segmented according to the color features of the warning signs (such as red, yellow, etc.) to extract the regions that may contain warning signs. The transmission line device features include conductor features and tower structure features, etc. For conductor feature extraction, the LSD (Line Segment Detector) line detection algorithm is used to detect the conductors. The LSD algorithm can quickly and accurately detect the straight line segments in the image and is suitable for the detection of conductors in transmission lines. The extracted different types of inherent image features are fused to form a comprehensive feature vector. For example, for substation devices, the device contour, insulator features, and warning sign features can be spliced to obtain a feature vector containing various information. The YOLOv5 object detection model and the ResNet50 classification model are fused. YOLOv5 can quickly and accurately detect the positions of target objects in the surveillance video frames, and ResNet50 can classify the detected targets to judge their states (the operating states of power equipment, and the specific content can be obtained through image analysis).

[0030] In some embodiments of the present application, before identifying the attributes and dynamic information of moving objects in the real-time surveillance video, the method further includes, Search for multiple fixed reference objects around the target power equipment, determine the actual position coordinates of the fixed reference objects and the pixel position coordinates of the fixed reference objects in the surveillance video, and determine a coordinate conversion scale ratio. The pixel position coordinates in the surveillance video are converted into actual position coordinates through the coordinate conversion scale ratio.

[0031] In this embodiment, before adopting a multi-target tracking system in the real-time monitoring and risk analysis stage, it is extremely crucial to calibrate the distance between the target power equipment and the moving objects in the monitoring video. The picture presented by the monitoring video is two-dimensional, and the position and distance information of the moving objects are based on pixel coordinates, rather than the actual spatial position and distance. Without calibration, subsequent analysis and tracking based on these uncalibrated data will lead to large errors. For example, when judging whether a moving object is approaching the target power equipment and evaluating the possible risks it may pose to the equipment, inaccurate distance information will make the risk analysis lose reliability, thereby affecting the effectiveness of the entire monitoring system. Select objects with stable physical structures and not easily movable as fixed references, such as permanent buildings (such as utility poles, distribution box enclosures, etc.) around power facilities, and solid ground markings (such as special concrete marker blocks). These objects can maintain their positions unchanged for a long time to ensure the accuracy of subsequent coordinate conversions. The reference objects should have obvious features in the monitoring video for easy and accurate identification and positioning. For example, they have unique shapes (such as triangular warning signs), bright colors (such as red identification posts), or specific textures (such as floor tiles with specific patterns).

[0032] In this embodiment, computer vision techniques, such as algorithms like template matching and feature point detection, are used to accurately identify each fixed reference object in the monitoring video. The template matching algorithm can search and match in the video frames by pre-making an image template of the reference object to find the position of the reference object. Feature point detection algorithms (such as SIFT, SURF, etc.) can extract unique feature points of the reference object and determine the position of the reference object in the video through the matching of feature points. Once the position of the reference object in the video is identified, its pixel coordinates in the video frame can be determined. Usually, with the upper left corner of the video frame as the origin, the horizontal direction as the X-axis, and the vertical direction as the Y-axis, measure the pixel coordinate values of the center point or specific feature points of the reference object and record them as the pixel position coordinates of the reference object in the monitoring video.

[0033] Step S103, obtain the real-time monitoring video around the target power equipment, identify the attributes and dynamic information of the moving objects in the real-time monitoring video, and predict the movement information of the moving objects based on the attributes and dynamic information of the moving objects.

[0034] In this embodiment, the attributes of the moving objects are identified, and the attributes include the category of the moving objects (such as people, vehicles, etc.) and the contour size (external structure situation), and the dynamic information includes the moving speed, moving trajectory, position, etc.

[0035] In some embodiments of the present application, identifying the attributes and dynamic information of the moving objects in the real-time monitoring video includes Identify the attributes and contour sizes of moving objects in the surveillance video through an identification model. The attributes of the moving objects include the categories of the moving objects and the contour sizes. The dynamic information of the moving objects includes the moving speed and the moving trajectory. The moving speed and the moving trajectory of the moving objects in the surveillance video are identified through the optical flow method and the Kalman filter.

[0036] In this embodiment, the optical flow method is adopted to estimate the displacement by analyzing the movement of pixels between adjacent frames. For people and vehicles, combined with the pixel-actual distance conversion ratio obtained by camera calibration, the pixel displacement is converted into the actual speed. For people, the Kalman filter algorithm is used to predict the position at the next moment according to the current position and speed information, and the trajectory is continuously tracked. The vehicle also uses the SORT algorithm to record the positions at different times and predicts the trajectory through the Kalman filter.

[0037] In some embodiments of the present application, predict the movement information of the moving object based on the attributes and dynamic information of the moving object, including, Input the attributes and dynamic information of the moving object within a current period of time into the moving object prediction model, and the moving object prediction model outputs the movement information of the moving object within a future period of time.

[0038] In this embodiment, collect a large amount of historical trajectory and speed data of moving objects, and analyze their movement laws in different scenarios and different time periods. For example, statistically analyze the average speed and common driving trajectories of vehicles on a certain section at different time periods. For a new moving object, predict its trajectory and speed according to the scene and time it is in, and refer to the historical data. Use time series analysis methods, such as ARIMA (Autoregressive Integrated Moving Average Model), to model the historical speed and trajectory data of the moving object, and predict the speed, trajectory and position within a future period of time.

[0039] Step S104, adjust the basic danger area of the target power equipment according to the movement information of the moving object and the current state of the target power equipment to obtain the danger area of the target power equipment. When the moving object is in the danger area, issue a warning to the periphery of the target power equipment, so as to achieve dynamic anti-electric shock warning.

[0040] In this embodiment, when the moving object is in the danger area (the adjusted area), issue anti-electric shock warnings such as corresponding sound and light to drive away or warn the moving object.

[0041] In some embodiments of the present application, adjust the basic danger area of the target power equipment according to the movement information of the moving object and the current state of the target power equipment to obtain the danger area of the target power equipment, including, Calculate the degree of intersection between the moving object and the basic dangerous area of the target power equipment according to the movement information of the moving object, and adjust the basic dangerous area of the target power equipment according to the degree of intersection and the current state of the target power equipment to obtain the dangerous area of the target power equipment.

[0042] In this embodiment, the movement information includes trajectory, speed, and position. The predicted time period is until the moving object leaves the peripheral area of the target power equipment. Evaluate the overlapping length or area of the trajectory of the moving object in the basic dangerous area, evaluate the walking angle of the moving object, and obtain the first intersection index, the second intersection index, and the third intersection index by separately considering the overlapping length or area of the trajectory in the basic dangerous area, the walking angle, and the speed. Determine the degree of intersection based on the first intersection index, the second intersection index, and the third intersection index. The specific formula is as follows: ; Wherein, is the degree of intersection, , , are the influence weights of the first intersection index, the second intersection index, and the third intersection index respectively, , , are the first intersection index, the second intersection index, and the third intersection index respectively, is , , the maximum value among the three, is a preset constant, represents the correction of the sum of the three by the maximum value among the three. Different degrees of intersection correspond to different adjustment coefficients, and the basic dangerous area is corrected by the basic dangerous area * (1 + adjustment coefficient).

[0043] Correspondingly, the present application also provides an anti-electric shock warning system based on monitoring video network data. As Figure 2 shown, it includes The first module is used to determine the position of the target power equipment, collect the traffic flow information and the history of electric shock accidents near the target power equipment, and divide the basic dangerous area of the target power equipment accordingly; The second module is used to screen out the inherent image features of the target power equipment in the monitoring video, establish a template of the target power equipment based on the inherent image features, and identify the target power equipment and its current state in the monitoring video through the template of the target power equipment; A third module is configured to obtain a real-time monitoring video around a target power device, identify the attributes and dynamic information of moving objects in the real-time monitoring video, and predict the movement information of the moving objects based on the attributes and dynamic information of the moving objects; A fourth module is configured to adjust the basic danger area of the target power device according to the movement information of the moving object and the current state of the target power device to obtain the danger area of the target power device, and issue a warning around the target power device when the moving object is in the danger area, so as to achieve dynamic anti-electric shock warning.

[0044] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Divide the basic danger area of the target power device according to the traffic flow information and electric shock historical accident information near the target power device, consider the past traffic flow and accident occurrence of the target power device, determine the size of a basic danger area, and provide a reliable basis for the subsequent warning of the power device. Establish a template of the target power device according to the inherent image features, so as to identify the current state of the target power device in video monitoring, and facilitate the subsequent adjustment of the warning range.

[0045] 2. Identify and predict the movement information of moving objects around the target power device to predict the situation of moving objects entering the basic danger area, and adjust the basic danger area of the target power device according to the movement information of the moving object and the current state of the target power device, so as to achieve adaptive early warning and warning, realize dynamic anti-electric shock warning and automatic anti-electric shock warning, improve the accuracy and adaptability of the anti-electric shock warning of the power device, reduce the probability of warning errors, and ensure the safety of pedestrians and vehicles and the normal operation of the power device.

[0046] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation scenarios of the present invention.

[0047] Those skilled in the art can understand that the drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the drawings are not necessarily essential for implementing the present invention.

[0048] Those skilled in the art can understand that the modules in the system in the implementation scenario can be distributed in the system of the implementation scenario according to the description of the implementation scenario, or can be correspondingly changed and located in one or more systems different from this implementation scenario. The modules in the above implementation scenario can be combined into one module, or can be further split into multiple sub-modules.

[0049] As mentioned above, the above are only the preferred specific implementation manners of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered within the protection scope of the present invention.

Claims

1. The anti-electric shock warning method based on monitoring video network data is characterized in that: include, Determine the location of the target power equipment, and collect traffic flow information and historical electric shock accident information near the target power equipment to divide the basic danger zone of the target power equipment; Filter out the inherent image features of the target power equipment in the monitoring video, establish a template of the target power equipment based on the inherent image features, and identify the target power equipment and its current state in the monitoring video through the template of the target power equipment; Acquire real-time monitoring video around the target power equipment, identify the attributes and dynamic information of the moving object in the real-time monitoring video, and predict the movement information of the moving object based on the attributes and dynamic information of the moving object; The basic danger zone of the target power equipment is adjusted according to the movement information of the moving object and the current state of the target power equipment to obtain the danger zone of the target power equipment. When the moving object is in the danger zone, a warning is issued to the surrounding area of ​​the target power equipment to achieve dynamic warning against electric shock.

2. The anti-electric shock warning method based on monitoring video network data according to claim 1 is characterized in that: The traffic flow information and historical electric shock accident information near the target power equipment are collected to divide the basic danger area of ​​the target power equipment. include, Collect traffic flow information and historical electric shock accident information within a preset range near the target power equipment, construct a traffic flow distribution map and a historical electric shock accident distribution map respectively, align the traffic flow distribution map and the historical electric shock accident distribution map, and superimpose the traffic flow distribution map and the historical electric shock accident distribution map to generate a traffic flow-electric shock accident distribution map, and the corresponding positions on the traffic flow-electric shock accident distribution map are marked with traffic flow characteristics and electric shock accident event parameters; The distribution degree of electric shock accidents is determined by the position of the electric shock accident event parameters on the traffic flow-electric shock accident distribution map, and the preset range near the target power equipment is gridded according to the distribution degree of electric shock accidents, and the area near the target power equipment is divided into two categories: the area where electric shock has occurred and the area where electric shock has not occurred; Calculate the distances from the area where electric shock has occurred and the area where electric shock has not occurred to the target power equipment; The regional risk levels of the area where electric shock has occurred and the area where electric shock has not occurred are determined respectively, and the basic hazardous area of ​​the target power equipment is determined by combining the regional risk levels of the area where electric shock has occurred and the area where electric shock has not occurred.

3. The anti-electric shock warning method based on monitoring video network data according to claim 2 is characterized in that: Determine the regional risk levels for areas where electric shock has occurred and areas where electric shock has not occurred, including: For the area where electric shock has occurred, confirm the average value of the traffic flow characteristics of the grid under the area where electric shock has occurred and the adjacent grids, calculate the ratio of the area of ​​the grid under the area where electric shock has occurred to the area within a preset range near the target power equipment, and determine the regional risk level of the area where electric shock has occurred based on the average value of the traffic flow characteristics, the distance between the area where electric shock has occurred and the target power equipment, the electric shock accident event parameters under the area where electric shock has occurred, and the ratio of the area of ​​the grid under the area where electric shock has occurred to the area within a preset range near the target power equipment; For areas where electric shock has not occurred, the distance between the area where electric shock has not occurred and the nearest adjacent area where electric shock has occurred is calculated, recorded as the nearest distance, and the ratio of the grid area under the area where electric shock has not occurred to the area within a preset range near the target power equipment is calculated. The regional risk level of the area where electric shock has not occurred is determined based on the traffic flow characteristics of the area where electric shock has not occurred, the nearest distance, and the distance between the area where electric shock has not occurred and the target power equipment.

4. The anti-electric shock warning method based on monitoring video network data according to claim 1 is characterized in that: Filter out the inherent image features of the target power equipment in the surveillance video, and establish a template of the target power equipment based on the inherent image features, including: Extract image features from the monitoring video according to the structure of the target power equipment and the identifiers of the target power equipment, fuse different types of image features, and generate a fused feature vector; A classification model is constructed according to the different operating states of the target power equipment, and a template of the target power equipment is established based on the fused feature vector and the classification model.

5. The anti-electric shock warning method based on monitoring video network data according to claim 1 is characterized in that: Before identifying the attributes and dynamic information of the moving object in the real-time monitoring video, the method further includes: A plurality of fixed reference objects are searched around the target power equipment, the actual position coordinates of the fixed reference objects and the pixel position coordinates of the fixed reference objects in the monitoring video are determined, and a coordinate conversion scale is determined, and the pixel position coordinates in the monitoring video are converted into the actual position coordinates through the coordinate conversion scale.

6. The anti-electric shock warning method based on monitoring video network data according to claim 1 is characterized in that: Identify the properties and dynamic information of moving objects in real-time surveillance video, including: The attributes and outline size of the moving object in the surveillance video are identified by the recognition model, and the attributes of the moving object include the category and outline size of the moving object; The dynamic information of the moving object includes the moving speed and moving trajectory. The moving speed and moving trajectory of the moving object in the surveillance video are identified by optical flow method and Kalman filtering.

7. The method for preventing electric shock based on monitoring video network data according to claim 6, characterized in that: Predicting movement information of the moving object based on the properties and dynamic information of the moving object, including: The attributes and dynamic information of the mobile objects in the current period of time are input into the mobile object prediction model, and the mobile object prediction model outputs the movement information of the mobile objects in the future period of time.

8. The method for preventing electric shock based on monitoring video network data according to claim 7, characterized in that: According to the movement information of the moving object and the current state of the target power equipment, the basic danger zone of the target power equipment is adjusted to obtain the danger zone of the target power equipment. include, The intersection degree between the mobile object and the basic dangerous area of ​​the target power equipment is calculated according to the movement information of the mobile object, and the basic dangerous area of ​​the target power equipment is adjusted according to the intersection degree and the current state of the target power equipment to obtain the dangerous area of ​​the target power equipment. 9.An anti-electric shock warning system based on monitoring video network data, characterized in that: include, The first module is used to determine the location of the target power equipment and collect traffic flow information and historical electric shock accident information near the target power equipment, so as to divide the basic danger zone of the target power equipment; The second module is used to screen out the inherent image features of the target power equipment in the monitoring video, establish a template of the target power equipment based on the inherent image features, and identify the target power equipment and the current state in the monitoring video through the template of the target power equipment; The third module is used to obtain real-time monitoring video around the target power equipment, identify the attributes and dynamic information of the mobile object in the real-time monitoring video, and predict the movement information of the mobile object based on the attributes and dynamic information of the mobile object; The fourth module is used to adjust the basic danger zone of the target power equipment according to the movement information of the mobile object and the current state of the target power equipment to obtain the danger zone of the target power equipment. When the mobile object is in the danger zone, a warning is issued to the surrounding area of ​​the target power equipment to achieve dynamic anti-electric shock warning.

Images