Substation operation risk identification method based on multi-view video and high-precision positioning

Abstract

The invention relates to a substation operation risk identification method based on a multi-view video and high-precision positioning. Acquiring video data of a working site through a plurality of cameras with fixed visual angles and mobile video acquisition equipment; a high-precision positioning system is used for obtaining three-dimensional space coordinates of operators and equipment in real time; establishing a three-dimensional digital twinborn model of the substation equipment, and performing dynamic scene reconstruction based on the multi-view video stream to generate a real-time three-dimensional scene of the operation site; fusing the positioning data and the three-dimensional scene by adopting a space-time fusion algorithm to generate a dynamic digital portrait of the operator; and carrying out real-time analysis on behaviors and positions of operators by using a risk assessment algorithm based on a preset risk rule, calculating to obtain a risk assessment value, and setting a feedback mechanism to continuously optimize positioning and scene reconstruction precision. According to the invention, efficient, accurate and real-time identification and early warning of the operation risk of the transformer substation are realized, and the safety management level of an operation site is effectively improved.

Description

Technical Field

[0001] This invention relates to the technical field of substation operation monitoring, and more particularly to a method for identifying substation operation risks based on multi-view video and high-precision positioning. Background Technology

[0002] In substation operations, traditional risk identification methods primarily rely on manual observation and experience-based judgment. However, manual observation is susceptible to factors such as personnel fatigue and lack of concentration, exhibiting intermittent and subjective issues, making it difficult to provide continuous, objective, and accurate monitoring of the work site. Experience-based judgment, on the other hand, lacks quantitative basis and struggles to accurately assess the degree of risk in complex substation environments and operational procedures. Furthermore, some relatively simple monitoring methods, such as single-point displacement sensors or infrared detection, only provide limited, localized information and cannot comprehensively reflect the complexities of the work site.

[0003] Furthermore, traditional methods have monitoring blind spots in areas obstructed by equipment (such as behind transformers, gaps in switchgear, etc.), making it difficult to effectively identify and warn of potential risks. Regarding positioning accuracy, traditional algorithms primarily rely on two-dimensional image analysis, making it difficult to accurately quantify the spatial distance between workers and live equipment. This results in inaccurate and untimely risk assessments, failing to meet the high-precision, high-reliability risk identification requirements of substation operations.

[0004] Therefore, it is essential for those skilled in the art to design a technology that can effectively identify risks through comprehensive monitoring. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a substation operation risk identification method based on multi-view video and high-precision positioning. This method overcomes the problems of blind spots and low positioning accuracy inherent in traditional monitoring methods, enabling efficient, accurate, and real-time identification and early warning of substation operation risks, and effectively improving the safety management level of the work site.

[0006] To achieve the above objectives, the technical solution adopted by this invention is: to provide a substation operation risk identification method based on multi-view video and high-precision positioning, comprising the following steps: Multiple fixed-view cameras and mobile video acquisition devices deployed in the substation work area are used to simultaneously acquire multiple channels of on-site work video. By having workers wear name tags with positioning functions, the two-dimensional coordinates of the workers are determined. Combined with the distance between the workers and multiple monitoring devices in multiple on-site video feeds, the three-dimensional spatial coordinates of the workers and equipment are determined. A three-dimensional digital twin model of the substation equipment is established, and dynamic scene reconstruction is performed based on the multi-channel on-site operation video to generate a real-time three-dimensional scene of the operation site. The three-dimensional spatial coordinates of the workers and equipment are spatiotemporally fused with the real-time three-dimensional scene of the work site using a spatiotemporal fusion algorithm to generate a real-time dynamic digital portrait of the workers. Based on preset risk rules, including safe distance thresholds, equipment restricted areas, and high-risk action modes, a risk assessment value is obtained in real time by using a risk assessment algorithm to calculate the behavior and location of the operator. The risk assessment value is used to determine whether the operator is currently engaging in risky work behavior. When the risk assessment value is greater than the preset threshold, it is determined that the operator has engaged in risky work behavior, including work behavior that is risky or the location is in a dangerous area, and multi-level early warning signals are generated and pushed to the monitoring terminal.

[0007] Preferably, the method for constructing the three-dimensional digital twin model of the substation equipment includes precise modeling of the geometry, size, location, and electrical connections of the substation equipment to ensure consistency with the actual substation equipment.

[0008] Preferably, in the three-dimensional digital twin model of the substation equipment: The three-dimensional spatial coordinates of the operators and equipment are transformed into the three-dimensional scene coordinate system where the three-dimensional digital twin model of the substation equipment is located. Using the known substation equipment location information and the coordinate system parameters of the positioning system, the rotation matrix and translation vector from the positioning system coordinate system to the scene coordinate system are calculated, and the coordinate transformation is performed on each coordinate point in the three-dimensional spatial coordinate system.

[0009] Preferably, the spatiotemporal fusion algorithm specifically includes the following processing: Establish a unified time coordinate system by adding timestamps to each data point in the two-dimensional and three-dimensional spatial coordinates, and using the video frame rate of the three-dimensional spatial coordinates as a reference, perform time interpolation on the two-dimensional coordinates to align the two in the time dimension. For each time point, the position of the operator in the 3D scene is calculated based on the transformed coordinates, and combined with the equipment information in the scene, a dynamic digital profile containing the operator's position, posture, and relative positional relationship with the surrounding equipment is generated.

[0010] Even better, the Kalman filter algorithm is used to estimate and correct uncertainties and errors in the spatiotemporal fusion process. By predicting and updating, the position and working status of the operators are continuously optimized, improving the fusion accuracy and stability. Specifically, this includes: Predict the current working status based on the working status of the previous moment; The dynamic digital profile is corrected using the actual measurement data at the current moment, resulting in a more accurate fusion result.

[0011] Preferably, the implementation steps of the risk assessment algorithm include: Risk factor extraction: Real-time extraction of workers' work behavior characteristics and location information from dynamic digital profiles; Safety distance assessment: For each worker in the scenario, calculate the distance between them and the surrounding live equipment, compare the calculated distance with the preset safety distance threshold, and if the actual distance is less than the threshold, it is determined that there is a safety distance risk. The degree of risk is quantified according to the extent to which the distance exceeds the threshold. Restricted area determination: Based on the layout of substation equipment and safety specifications, the restricted area of ​​the equipment is delineated in advance in the three-dimensional digital twin model. Whether the operators enter the restricted area is detected in real time. If they do enter, it is determined that there is a risk violation, which is recorded, archived and assigned with corresponding risk weights. High-risk action recognition: Using pose estimation and action recognition algorithms in video AI technology, the action sequence of workers is analyzed. Through a trained deep learning model, it is possible to identify whether there are high-risk action patterns such as climbing or approaching live parts. Once a high-risk action is identified, it is determined that there is a risky behavior and the corresponding risk warning is triggered. Comprehensive risk assessment: A weighted summation method is used to comprehensively calculate the risks of safe distance, violation of restricted areas, and high-risk actions to obtain the final risk assessment value. The weights of different risk factors are dynamically adjusted according to the importance and degree of danger of the actual on-site operation.

[0012] More preferably, the formula for calculating the risk assessment value is as follows: Where R represents the risk assessment value, , and These represent the weights of the risks associated with safe distances, violations of restricted areas, and high-risk actions, respectively, to meet the following requirements. , This represents the risk value of the safe distance, calculated as the ratio of the actual distance to the safe distance threshold. This represents the risk value for violations of restricted areas. The value is 1 when workers enter the restricted area, and 0 otherwise. This represents the risk value of a high-risk action. The value is 1 when a high-risk action is identified, and 0 otherwise. When the risk assessment value exceeds the preset risk threshold, an early warning signal is triggered. At the same time, relevant information about the risk event is recorded, including the time, location, personnel involved, and risk type. The risk event is then fed back to the monitoring terminal, and management personnel take appropriate measures in a timely manner.

[0013] As an optimal solution, a feedback mechanism is also set up. Based on the movement trajectory and posture change information of the workers in multiple operation site videos, error analysis and compensation are performed on the acquired three-dimensional spatial coordinates. An adaptive filtering algorithm is used to correct the positioning data in real time, continuously improving the positioning results to obtain more accurate positioning results.

[0014] The beneficial effects of this invention are as follows: This invention provides a substation operation risk identification method based on multi-view video and high-precision positioning, which can effectively solve the shortcomings of traditional technologies. First, by using multi-view cameras and mobile video acquisition equipment, it achieves all-round, real-time monitoring of the substation operation site without blind spots, effectively eliminating equipment obstruction blind spots and ensuring the integrity of monitoring information. Second, by using a high-precision positioning system to obtain the three-dimensional spatial coordinates of operators and equipment in real time, combined with a three-dimensional digital twin model and spatiotemporal fusion algorithm, it achieves accurate three-dimensional reconstruction of the operation site and dynamic digital portraits of personnel, significantly improving the quantitative accuracy of the spatial relationship between personnel and equipment and overcoming the limitations of traditional two-dimensional image analysis in distance judgment. Third, based on preset risk rules and risk assessment algorithms, it can analyze the behavior and location of operators in real time, quickly and accurately identify risks such as violations of safe distances, intrusion into restricted areas, and high-risk actions, and realize multi-level early warning functions, providing timely and effective risk warnings for operators and managers, and enhancing on-site emergency response capabilities. Finally, through a feedback correction mechanism, the positioning data and scene reconstruction results are continuously optimized, improving the system's adaptability and stability to complex operating environments and ensuring the long-term reliability of risk identification. In summary, this invention significantly improves the efficiency, accuracy, and intelligence of substation operation risk identification, effectively ensuring the safety of operators and the stable operation of substations. Attached Figure Description

[0015] Figure 1 This is a flowchart of the substation operation risk identification method based on multi-view video and high-precision positioning of the present invention. Detailed Implementation

[0016] Please see Figure 1 As shown, this invention provides a substation operation risk identification method based on multi-view video and high-precision positioning, including the following steps: Multiple fixed-view cameras and mobile video acquisition devices deployed in the substation work area are used to simultaneously acquire multiple channels of on-site work video. In this embodiment, taking a 220kV substation work area as an example, multiple fixed-view cameras are deployed, installed on pillars around the substation at a height of approximately 3 meters, to ensure coverage of different angles of the entire work area. Simultaneously, mobile video acquisition devices with real-time transmission capabilities are provided to the workers, mounted on their safety helmets, to obtain on-site video from the workers' first-person perspective.

[0017] In addition, a high-precision positioning system, including positioning base stations and positioning tags, is installed inside the substation. The positioning base stations are fixed at multiple known locations within the substation, forming a positioning network; workers wear positioning tags and send wireless signals to the positioning base stations in real time to obtain their own three-dimensional spatial coordinates.

[0018] Fixed-view cameras and mobile video acquisition devices are activated simultaneously, acquiring multiple video feeds from the work site at a rate of 30 frames per second. This data includes operator procedures, equipment operating status, and environmental conditions. The video data is then transmitted in real-time via a wireless network to a data acquisition server on-site for storage and preprocessing.

[0019] The high-precision positioning system collects three-dimensional spatial coordinate data of personnel and equipment at a frequency of 10 times per second. The positioning base station receives signals sent by the positioning tags and uses ultra-wideband (UWB) positioning technology to calculate and record the coordinate position (x, y, z) of each positioning tag in three-dimensional space, while transmitting the data to the data acquisition server.

[0020] By having workers wear name tags with positioning functions, the two-dimensional coordinates of the workers are determined. Combined with the distance between the workers and multiple monitoring devices in multiple on-site video feeds, the three-dimensional spatial coordinates of the workers and equipment are determined. A three-dimensional digital twin model of the substation equipment is established, and dynamic scene reconstruction is performed based on the multi-channel on-site operation video to generate a real-time three-dimensional scene of the operation site. Based on the substation's design drawings and equipment parameters, a three-dimensional digital twin model of the substation equipment is constructed in computer software. This model accurately describes the equipment's geometry, dimensions, location, and electrical connections, ensuring a high degree of consistency with the actual substation equipment. For example, the main transformer model is set to have a length, width, and height of 3.5 meters, 2 meters, and 2.2 meters respectively, accurately simulating its appearance and location.

[0021] Dynamic scene reconstruction is performed using multi-view video streams from a data acquisition server and computer vision algorithms. First, feature extraction and matching are performed on video frames from multiple perspectives to identify common feature points in the video. Then, based on these feature points, algorithms such as bundle adjustment are used to calculate the camera's intrinsic and extrinsic parameters, thereby reconstructing the scene's 3D structure. Simultaneously, a 3D digital twin model is registered and fused with the reconstructed 3D scene to generate a real-time 3D scene of the work site, including equipment and personnel. Throughout this process, the position and posture information of the personnel in the 3D scene are updated in real-time based on feature matching and depth information from the video, ensuring a high degree of consistency between the 3D scene and the actual work site.

[0022] The three-dimensional spatial coordinates of the workers and equipment are spatiotemporally fused with the real-time three-dimensional scene of the work site using a spatiotemporal fusion algorithm to generate a real-time dynamic digital portrait of the workers. The spatiotemporal fusion algorithm specifically includes the following processing: Establish a unified time coordinate system by adding timestamps to each data point in the two-dimensional and three-dimensional spatial coordinates, and using the video frame rate of the three-dimensional spatial coordinates as a reference, perform time interpolation on the two-dimensional coordinates to align the two in the time dimension. For each time point, the position of the worker in the 3D scene is calculated based on the transformed coordinates. Combined with equipment information in the scene, a dynamic digital profile is generated, including the worker's position, posture, and relative positional relationship with surrounding equipment. Specifically, this includes: Time synchronization: Precise timestamps are added to the 3D spatial coordinate data and multi-view video data acquired by the high-precision positioning system. Based on the video frame rate, time interpolation is performed on the positioning data to align it with the video data in the time dimension.

[0023] Coordinate Transformation: Based on the known camera position and attitude information, and the coordinate system parameters of the positioning system, calculate the rotation matrix and translation vector from the positioning system coordinate system to the 3D scene coordinate system. Transform the 3D spatial coordinates in the positioning data into the 3D scene coordinate system using coordinate transformation formulas.

[0024] Feature matching: Extracting feature points of equipment and workers from the 3D scene, such as edges and corners of equipment and joints of workers. Simultaneously, predicting the possible location range of workers in the 3D scene based on positioning data narrows the search area for feature matching, improving matching efficiency.

[0025] Fusion computing: This involves fusing time-aligned and coordinate-transformed positioning data with 3D scene data. For each time point, the precise location of the worker in the 3D scene is determined based on the transformed coordinates. Combined with equipment information in the scene, a dynamic digital profile is generated, including the worker's position, posture, and relative positional relationship with surrounding equipment. For example, the fused data clearly shows the distance between the worker and live equipment, and whether the worker's movements and postures comply with safety regulations.

[0026] The Kalman filter algorithm is employed to estimate and correct uncertainties and errors in the spatiotemporal fusion process. Through prediction and updates, the system continuously optimizes the position and working status of personnel, improving fusion accuracy and stability. Specifically, this includes: Predict the current working status based on the working status of the previous moment; The dynamic digital profile is corrected using actual measurement data at the current moment, resulting in a more accurate fusion result, as detailed below: First, initialize the state vector and covariance matrix of the Kalman filter. The state vector contains estimates of the operator's position and velocity, and the covariance matrix represents the uncertainty of these estimates.

[0027] Prediction Steps The state at the current time step is predicted using the state transition model and the state estimate from the previous time step. The prediction formula is: in, It is the prior state estimate at the current moment, and F is the state transition matrix. It is the posterior state estimate of the previous time step.

[0028] Simultaneously, update the prior estimated covariance matrix: in, Q is the prior estimate covariance matrix, and Q is the process noise covariance matrix.

[0029] Update steps: Compare the current measurement data (such as 3D coordinates from a high-precision positioning system and attitude estimation from video analytics) with the predicted state, and adjust the predicted state using Kalman gain to obtain a more accurate estimate. Calculate Kalman gain: in, H is the Kalman gain, H is the observation matrix, and R is the measurement noise covariance matrix.

[0030] The state estimate updated using Kalman gain is expressed as follows: ( ) in, It is the posterior state estimate at the current moment. It is the measurement value at the current moment.

[0031] The updated posterior estimated covariance matrix is ​​as follows: Where I is the identity matrix.

[0032] Kalman filtering continuously adjusts the state estimate and covariance matrix through iterative prediction and update steps, reducing estimation errors. In substation operation risk assessment, this process is repeated to ensure the accuracy of personnel position and attitude estimation, thereby enhancing the reliability of risk assessment. For example, when personnel approach live equipment, Kalman filtering can promptly and accurately reflect this change, enabling the risk assessment system to react quickly.

[0033] Based on preset risk rules, including safe distance thresholds, equipment restricted areas, and high-risk action modes, a risk assessment value is obtained in real time by using a risk assessment algorithm to calculate the behavior and location of the operator. The risk assessment value is used to determine whether the operator is currently engaging in risky work behavior. The preset risk rules are as follows: Extract the behavioral characteristics and location information of the workers from the real-time 3D scene, including their posture (such as the angle of arm extension, the degree of leg bending, etc.) and their relative position to the equipment (such as the distance to the live equipment, whether they are within the operating range of the equipment, etc.).

[0034] According to the safety regulations for substations, a safe distance threshold for live equipment is set. For example, for 220kV live equipment, the safe distance threshold is set to 1.8 meters. The Euclidean distance between the operator and the live equipment is calculated in real time. If the actual distance is less than the threshold, a safety distance risk is determined, and the safety distance risk value is calculated according to a certain quantification rule based on the degree to which the distance exceeds the threshold. For example, when the actual distance is 1.5 meters, the safety distance risk value can be calculated as (1.8 - 1.5) / 1.8 = 0.167. The specific quantification rule can be determined according to the actual situation.

[0035] In the 3D digital twin model, the restricted area of ​​the equipment is pre-defined, such as setting a circular restricted area with a radius of 1 meter around high-voltage equipment. The system monitors in real time whether workers enter the restricted area; if they do, the violation is recorded and the risk value for the restricted area violation is set to 1; otherwise, it is set to 0.

[0036] This study utilizes deep learning-based video AI technology to analyze the movement sequences of workers. A convolutional neural network (CNN) model is trained using a large amount of video data labeled with high-risk actions (such as climbing and approaching live parts). In practical applications, real-time video frames of workers are input into the trained CNN model to identify high-risk movement patterns. Once a high-risk action is identified, a corresponding risk warning is triggered, and the risk value of the high-risk action is set to 1; otherwise, it is set to 0.

[0037] Comprehensive Risk Assessment: The weights for safety distance risk, restricted area violation risk, and high-risk action risk are set at 0.5, 0.3, and 0.2, respectively. The actual weights can be dynamically adjusted based on the importance and hazard level of the on-site operation. The risk assessment value is calculated using the following formula: Where R represents the risk assessment value, , and These represent the weights of the risks associated with safe distances, violations of restricted areas, and high-risk actions, respectively, to meet the following requirements. , This represents the risk value of the safe distance, calculated as the ratio of the actual distance to the safe distance threshold. This represents the risk value for violations of restricted areas. The value is 1 when workers enter the restricted area, and 0 otherwise. This represents the risk value of a high-risk action. The value is 1 when a high-risk action is identified, and 0 otherwise. For example, when the safe distance risk value is 0.167, the restricted area violation risk value is 0, and the high-risk action risk value is 1, the risk assessment values ​​are as follows: R=0.5×0.167+0.3×0+0.2×1=0.233+0+0.2=0.433; When the risk assessment value exceeds the preset risk threshold, an early warning signal is triggered. At the same time, relevant information about the risk event is recorded, including the time, location, personnel involved, and risk type. The risk event is then fed back to the monitoring terminal, and management personnel take appropriate measures in a timely manner.

[0038] In this embodiment, the preset risk threshold is 0.5. When R=0.433, it is less than the preset risk threshold, and no warning signal is triggered; when the calculated risk assessment value exceeds the threshold, a warning signal is triggered immediately.

[0039] When the risk assessment value is greater than the preset threshold, it is determined that the operator has engaged in risky work behavior, including work behavior that is risky or the location is in a dangerous area, and multi-level early warning signals are generated and pushed to the monitoring terminal.

[0040] In addition, the early warning signals are divided into multiple levels, graded according to the magnitude of the risk assessment value. For example, a risk assessment value between 0.5 and 0.7 is a Level 2 warning, reminding workers to pay attention to safety; a value above 0.7 is a Level 1 warning, urgently reminding workers to stop work and evacuate the danger zone. Simultaneously, relevant information about risk events (such as time, location, personnel involved, and risk type) is recorded and fed back to the monitoring terminal, enabling managers to promptly understand the on-site situation and take appropriate measures. Through the analysis of historical risk events, risk rules and algorithm parameters are continuously optimized to improve the accuracy and reliability of risk assessment.

[0041] In addition, the present invention also sets up a feedback mechanism, which performs error analysis and compensation on the acquired three-dimensional spatial coordinates based on the movement trajectory and posture change information of the workers in multiple operation site videos, and uses an adaptive filtering algorithm to correct the positioning data in real time, continuously repairing and obtaining more accurate positioning results.

[0042] The specific steps are as follows: Video Data and 3D Scene Correction: Multiple video feeds from the work site are compared with real-time 3D scenes. Using texture and edge features from the videos, a feature-matching-based optimization algorithm is employed to correct the geometry of the 3D scene. For example, if a discrepancy is found between the edge of a device in the video and the edge of its model in the 3D scene, the parameters of the device model in the 3D scene are adjusted based on the feature point matching results in the video, making it more closely match the actual shape and position of the device, thereby improving the accuracy of scene reconstruction.

[0043] Location data correction: Based on the motion trajectory and posture change information of workers in multi-view videos, error analysis is performed on the location data acquired by the high-precision positioning system. An adaptive filtering algorithm is used to estimate the errors in the location data in real time and to compensate and correct the data. For example, when a systematic deviation is found in the location data within a certain time period, the coordinate values ​​of the location data are adjusted according to the actual motion trajectory of the workers in the video, so that the corrected location data more accurately reflects the actual position of the workers. By continuously repeating this feedback correction process, the algorithm parameters are optimized, improving positioning accuracy and scene reconstruction effects, thereby enhancing the reliability and accuracy of the entire risk identification system.

[0044] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for identifying substation operation risks based on multi-view video and high-precision positioning, characterized in that, Includes the following steps: Multiple fixed-view cameras and mobile video acquisition devices deployed in the substation work area are used to simultaneously acquire multiple channels of on-site work video. By having workers wear name tags with positioning functions, the two-dimensional coordinates of the workers are determined. Combined with the distance between the workers and multiple monitoring devices in multiple on-site video feeds, the three-dimensional spatial coordinates of the workers and equipment are determined. A three-dimensional digital twin model of the substation equipment is established, and dynamic scene reconstruction is performed based on the multi-channel on-site operation video to generate a real-time three-dimensional scene of the operation site. The three-dimensional spatial coordinates of the workers and equipment are spatiotemporally fused with the real-time three-dimensional scene of the work site using a spatiotemporal fusion algorithm to generate a real-time dynamic digital portrait of the workers. Based on preset risk rules, including safe distance thresholds, equipment restricted areas, and high-risk action modes, a risk assessment value is obtained in real time by using a risk assessment algorithm to calculate the behavior and location of the operator. The risk assessment value is used to determine whether the operator is currently engaging in risky work behavior. When the risk assessment value is greater than the preset threshold, it is determined that the operator has engaged in risky work behavior, including work behavior that is risky or the location is in a dangerous area, and multi-level early warning signals are generated and pushed to the monitoring terminal.

2. The substation operation risk identification method based on multi-view video and high-precision positioning according to claim 1, characterized in that, The method for constructing the three-dimensional digital twin model of the substation equipment includes precise modeling of the geometry, size, location, and electrical connections of the substation equipment to ensure consistency with the actual substation equipment.

3. The substation operation risk identification method based on multi-view video and high-precision positioning according to claim 1, characterized in that, In the three-dimensional digital twin model of the substation equipment: The three-dimensional spatial coordinates of the operators and equipment are transformed into the three-dimensional scene coordinate system where the three-dimensional digital twin model of the substation equipment is located. Using the known substation equipment location information and the coordinate system parameters of the positioning system, the rotation matrix and translation vector from the positioning system coordinate system to the scene coordinate system are calculated, and the coordinate transformation is performed on each coordinate point in the three-dimensional spatial coordinate system.

4. The substation operation risk identification method based on multi-view video and high-precision positioning according to claim 1, characterized in that, The spatiotemporal fusion algorithm specifically includes the following processing: Establish a unified time coordinate system by adding timestamps to each data point in the two-dimensional and three-dimensional spatial coordinates, and using the video frame rate of the three-dimensional spatial coordinates as a reference, perform time interpolation on the two-dimensional coordinates to align the two in the time dimension. For each time point, the position of the operator in the 3D scene is calculated based on the transformed coordinates, and combined with the equipment information in the scene, a dynamic digital profile containing the operator's position, posture, and relative positional relationship with the surrounding equipment is generated.

5. The substation operation risk identification method based on multi-view video and high-precision positioning according to claim 4, characterized in that, The Kalman filter algorithm is used to estimate and correct uncertainties and errors in the spatiotemporal fusion process. Through prediction and updating, the algorithm continuously optimizes the position and working status of personnel, improving fusion accuracy and stability. Specifically, this includes: Predict the current working status based on the working status of the previous moment; The dynamic digital profile is corrected using the actual measurement data at the current moment, resulting in a more accurate fusion result.

6. The substation operation risk identification method based on multi-view video and high-precision positioning according to claim 1, characterized in that, The implementation steps of the risk assessment algorithm include: Risk factor extraction: Real-time extraction of workers' work behavior characteristics and location information from dynamic digital profiles; Safety distance assessment: For each worker in the scenario, calculate the distance between them and the surrounding live equipment, compare the calculated distance with the preset safety distance threshold, and if the actual distance is less than the threshold, it is determined that there is a safety distance risk. The degree of risk is quantified according to the extent to which the distance exceeds the threshold. Restricted area determination: Based on the layout of substation equipment and safety specifications, the restricted area of ​​the equipment is delineated in advance in the three-dimensional digital twin model. Whether the operators enter the restricted area is detected in real time. If they do enter, it is determined that there is a risk violation, which is recorded, archived and assigned with corresponding risk weights. High-risk action recognition: Using pose estimation and action recognition algorithms in video AI technology, the action sequence of workers is analyzed. Through a trained deep learning model, it is possible to identify whether there are high-risk action patterns such as climbing or approaching live parts. Once a high-risk action is identified, it is determined that there is a risky behavior and the corresponding risk warning is triggered. Comprehensive risk assessment: A weighted summation method is used to comprehensively calculate the risks of safe distance, violation of restricted areas, and high-risk actions to obtain the final risk assessment value. The weights of different risk factors are dynamically adjusted according to the importance and degree of danger of the actual on-site operation.

7. The substation operation risk identification method based on multi-view video and high-precision positioning according to claim 6, characterized in that, The formula for calculating the risk assessment value is as follows: Where R represents the risk assessment value, , and These represent the weights of the risks associated with safe distances, violations of restricted areas, and high-risk actions, respectively, to meet the following requirements. , This represents the risk value of the safe distance, calculated as the ratio of the actual distance to the safe distance threshold. This represents the risk value for violations of restricted areas. The value is 1 when workers enter the restricted area, and 0 otherwise. This represents the risk value of a high-risk action. The value is 1 when a high-risk action is identified, and 0 otherwise. When the risk assessment value exceeds the preset risk threshold, an early warning signal is triggered. At the same time, relevant information about the risk event is recorded, including the time, location, personnel involved, and risk type. The risk event is then fed back to the monitoring terminal, and management personnel take appropriate measures in a timely manner.

8. The substation operation risk identification method based on multi-view video and high-precision positioning according to claim 1, characterized in that, A feedback mechanism was also set up to analyze and compensate for errors in the acquired three-dimensional spatial coordinates based on the movement trajectory and posture change information of the workers in multiple operation site videos. An adaptive filtering algorithm was used to correct the positioning data in real time, and continuous repair was carried out to obtain more accurate positioning results.

Images