Substation defect video analysis system

Abstract

The invention provides a transformer substation defect video analysis system. The system is provided with a hard disk video recorder, a camera ledger management module, a task management module, an algorithm analysis module, an alarm receiving and pushing module and a video downloading module, the machine account management module of the hard disk video recorder and the camera provides IP addresses, core drive types, channel numbers and corresponding user information of the hard disk video recorder and the camera, and generates a historical video RTSP address for subsequent defect analysis; the task management module aims at efficiently processing analysis tasks and creating new analysis tasks; the algorithm analysis module is responsible for identifying defects; the alarm module stores defect analysis data in a database and pushes the defect analysis data to a user interface in real time; the video downloading module can download defect videos to the local, so that subsequent maintenance and processing are facilitated; the system can perform defect analysis according to historical videos, generates corresponding alarm information, and improves the working efficiency of operation and maintenance personnel.

Description

Technical Field

[0001] The present invention relates to the technical field of video analysis and processing systems, and in particular to a substation defect video analysis system. Background Art

[0002] In the intelligent inspection of substations, a large number of cameras are installed inside the substation. These cameras can play a great role in the operation, safety and intelligent control of the substation. In the research on intelligent control systems, we are committed to maximizing the utilization of idle camera recording resources; when the system is idle, these recording resources can be used for in-depth image analysis, and real-time alarm prompts can be provided based on the specific defect types that maintenance personnel are concerned about; such intelligent analysis helps substation maintenance personnel to more effectively perform equipment maintenance and problem investigation, ensure that potential defects are discovered and repaired in time, thereby avoiding greater losses; in this way, not only the utilization rate of historical recording resources is improved, but also the need for repeated inspections of the same defect is reduced, and the video clips downloaded based on the defect can also be used for algorithm training, greatly improving the algorithm recognition accuracy; for this reason, the applicant proposed a substation defect video analysis system based on the above needs, which can perform defect analysis based on historical recordings, generate corresponding alarm information, and improve the work efficiency of operation and maintenance personnel. Summary of the Invention

[0003] In order to solve the above technical problems, the present invention proposes a substation defect video analysis system, which is provided by setting a hard disk recorder and camera ledger management module, a task management module, an algorithm analysis module, an alarm receiving and pushing module and a video downloading module; the hard disk recorder and camera ledger management module provides the IP address, core driver type, channel number and corresponding user information of the hard disk recorder and camera, generates the historical video RTSP address, and facilitates the subsequent defect analysis; the task management module is designed to efficiently process the analysis task and create new analysis tasks; the algorithm analysis module is responsible for identifying defects; the alarm module stores the defect analysis data in the database and pushes it to the user interface in real time; the video downloading module can download the defect video to the local computer, which is convenient for subsequent maintenance and processing; the system can perform defect analysis based on the historical video, generate corresponding alarm information, and improve the work efficiency of the operation and maintenance personnel.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] A substation defect video analysis system, characterized in that: the substation defect video analysis system is provided with a hard disk video recorder and camera ledger management module, a task management module, an algorithm analysis module, an alarm receiving and pushing module and a video download module; the hard disk video recorder and camera ledger management module mainly provides the IP address of the hard disk video recorder and the camera, the hard disk video recorder core driver type, the hard disk video recorder channel number to which the camera is connected, and the corresponding user name and password information, so as to generate a historical video RTSP address for the image algorithm analysis module to perform defect analysis on the video resources; the task management module can efficiently process analysis tasks, and users can create new analysis tasks according to specific needs and save them to the database. The task management module can select defect detection for camera videos within a specific time period, and users can choose to start analysis at leisure time or execute the analysis task immediately; when the analysis task reaches the predetermined start time, the background will automatically transfer the camera video resources to the algorithm analysis module for defect analysis ; Users can edit and delete analysis tasks. When an analysis task is in progress, users have the right to manually stop the analysis process; the algorithm analysis module is responsible for identifying defects; it supports frame-skipping and double-speed analysis of videos that may remain static for a long time to improve analysis efficiency; the analysis results are reported to the alarm receiving and pushing module through the MQTT protocol; the ceiling alarm module first stores the defect analysis data in the database, and then organizes the processed defect information into an easy-to-read format, and pushes it to the user interface in real time through WebSockets technology; the system can dynamically generate groups based on tasks, cameras and alarm content, allowing users to focus on observing the frequency and proportion of specific defects in the video; the video download module can download and save defect videos locally for subsequent maintenance and processing; the video download module also provides detailed video information, including the date, time and type of defect of the video, etc., to facilitate user classification management and quick retrieval; users can easily filter out the required video materials by timeline or defect type, greatly improving work efficiency.

[0006] Furthermore, the task management module of the substation defect video analysis system is configured with analysis speeds of 2x, 4x, 8x and 16x; the task management module can specify analysis defect types, including but not limited to: crossing the line, not wearing a safety helmet, not wearing prescribed work clothes, smoking, not using a safety rope, people falling, oil stains on the ground, suspended matter in the air, bird nests, damaged or missing covers, fire smoke, water accumulation on the indoor floor, small animal intrusion, wall cracks, broken insulators, equipment icing and broken silicone tubes, etc.

[0007] Furthermore, the task management module of the substation defect video analysis system has the function of avoiding overload of the algorithm analysis service; specifically:

[0008] Users can set the upper limit for the number of concurrent analysis tasks and the upper limit for the number of cameras being analyzed. When the number of analysis tasks reaches this limit, new tasks will be placed in a waiting queue. Once the scheduled execution time arrives and the task queue conditions are met, the newly added task will begin the analysis process and enter the analyzing state. The upper limit for analysis tasks is taskLimit, the upper limit for camera analysis is cameraLimit, the number of tasks being analyzed is analyzeTask, the number of cameras being analyzed is analyzeCamera, the number of cameras included in the task currently being started is waitCamera, and whether the task is executable is canRun:

[0009] 1) When analyzeTask=0, it means there is no task being analyzed; then canRun=true;

[0010] 2) When analyzeTask>0; analyzeTask+1<=taskLimit, and waitCamera+analyseCamera<=cameraLimit; then canRun=true, otherwise canRun=true;

[0011] When canRun=true, it means that the task can be started, otherwise the task will re-enter the pending analysis state and wait for the next task to be started; whenever a task is completed or the user manually terminates a task, an attempt will be made to start the pending analysis task with the shortest execution time, entering the above judgment logic; during task analysis, the user can choose to manually terminate the task, making it enter the manual termination state, or wait for the task to complete naturally, thus entering the analysis completion state; for tasks that are not in the analysis state, the user can also choose to manually delete them, which will simultaneously delete the analysis results and related analyzed files; clicking the Details button allows the user to view the options set when the task was created to fully understand the specific content of the task.

[0012] Furthermore, the alarm time calculation of the algorithm analysis module set in the substation defect video analysis system is as follows: let the alarm time be alarmTime, the video start time be videoStartTime, the number of frames in the video be frame, and the frame rate of the video be frameRate.

[0013]

[0014] By analyzing the number of frames where the defect occurs and the frame rate of the video, the specific time point (frame / frameRate) at which the defect occurs in the video is calculated. Based on this, the time of the defect alarm is the sum of the video start time and the duration of the defect location. This calculated alarm time helps locate the defect during subsequent video downloads. Users can adjust the download start and end times appropriately to download only the video clip containing the defect, quickly locating the problem without downloading the entire video file.

[0015] Furthermore, the analysis progress of the algorithm analysis module set in the substation defect video analysis system is specifically set as follows: the analysis module will regularly report the video analysis progress of a single camera every 5 seconds through the MQTT protocol; based on the task created by the user, involving multiple cameras, the system integrates the video analysis progress of all cameras and displays a comprehensive analysis progress to the user; let the overall progress be allProgress, the video analysis progress of the first camera be progress_1, the second one be progress_2, and so on to the progress of the nth camera progress_n; in this way, the analysis progress of the entire task is calculated and updated and pushed to the user in real time; the formula is:

[0016]

[0017] By summarizing the analysis progress of each camera, dividing it by the total number of cameras, and finally multiplying it by 100, the overall percentage progress of the recording task can be calculated. This calculation method ensures that users can understand the progress of the entire recording analysis task in real time, so as to better arrange and manage work. At the same time, the system will also display this overall progress through a user-friendly interface, allowing users to intuitively see the completion status of the task in real time.

[0018] Furthermore, the alarm receiving and pushing module set up in the substation defect video analysis system has the function of repeated alarms caused by the persistence of defects; specifically: when MQTT receives the defect message analyzed by the algorithm, it will immediately store the original analysis results in the database, and push them to the user interface in real time through WebSockets technology for rendering and display, ensuring low latency and high reliability of data transmission; in order to prevent repeated alarms due to the persistence of defects, the system will query the corresponding task result group groupId based on the task identifier taskId, camera identifier cameraId and alarm content alarmContent in the analysis result; if groupId can be queried based on taskId, cameraId and alarmContent, the analysis result will be classified into the corresponding task group, and the last alarm time of the group and the last analysis time of the task will be updated to the current analysis result If the groupId cannot be found, it means that the current analysis result has not been grouped. In this case, a new result group will be created, and the time of the most recent alarm and the time of the most recent analysis of the task will be recorded as the time corresponding to the current analysis result, and the number of alarms for the group will be initialized to 1. In the process of task group management, the system will also monitor the alarm frequency. If the number of alarms for a task group exceeds the set threshold within the preset time window, the system will automatically trigger a high-level alarm notification to remind the user to pay attention and deal with possible abnormal situations in a timely manner. In addition, the system will regularly analyze historical alarm data, identify common alarm patterns and potential risk points through machine learning algorithms, and provide data support for subsequent fault prevention and system optimization. Through this series of management measures, the system can more effectively reduce the interference of invalid alarms, improve the work efficiency of operation and maintenance personnel, and ensure the stability and reliability of the entire monitoring system.

[0019] Furthermore, the video downloading time of the video downloading module set in the substation defect video analysis system is set as follows: the alarm time is set to alarmTime, the video downloading start time is set to videoDownLoadStartTime, and the video downloading end time is set to videoDownLoadEndTime:

[0020] videoDownLoadStartTime=alarmTime-1

[0021] videoDownLoadEndTime=alarmTime+1

[0022] The default start and end times for video downloads are set to plus or minus 1 minute before and after the alarm time, simplifying the process of selecting the start and end times for video downloads. Users can preview and play the video and manually adjust the download start and end times to ensure that the video can be downloaded completely when the defect alarm occurs. This facilitates subsequent maintenance processing and allows users to use alarm videos for model training of analysis algorithms to further improve the accuracy of the analysis algorithms.

[0023] The benefits of this application are:

[0024] 1. The substation defect video analysis system can set the defect type and analysis speed according to needs, effectively improving the efficiency of video analysis and highlighting the key points;

[0025] 2. The substation defect video analysis system can view the system analysis status and progress in real time according to needs, meeting the user's needs to understand the system in a timely manner;

[0026] 3. The substation defect video analysis system can group and store the results according to the defect analysis, so that users can understand the content of the defect in a timely manner and download the defect video for detailed viewing and analysis;

[0027] 4. The substation defect video analysis system regularly analyzes historical alarm data and uses machine learning algorithms to identify common alarm patterns and potential risk points, providing data support for subsequent fault prevention and system optimization. Through this series of management measures, the system can more effectively reduce the interference of invalid alarms, improve the work efficiency of operation and maintenance personnel, and ensure the stability and reliability of the entire monitoring system. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a schematic diagram of the overall module connection and data flow of the present invention;

[0029] Figure 2 A schematic diagram showing a screenshot of a newly added task in the present invention;

[0030] Figure 3 This is a schematic diagram of the screenshot of the defect analysis results of the present invention;

[0031] Figure 4 This is a schematic diagram of downloading video clips of the time period where the defects of the present invention occur. DETAILED DESCRIPTION

[0032] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0033] like Figure 1-4 As shown, a substation defect video analysis system is shown. Figure 1As shown, the substation defect video analysis system is provided with a hard disk video recorder (NVR) and camera ledger management module, a task management module, an algorithm analysis module, an alarm receiving and pushing module and a video download module; the hard disk video recorder and camera ledger management module mainly provides the IP address of the hard disk video recorder and the camera, the hard disk video recorder core driver type, the hard disk video recorder channel number to which the camera is connected, and the corresponding user name and password information, so as to generate the historical video RTSP address for the image algorithm analysis module to perform defect analysis on the video resource; the task management module can efficiently process the analysis task, and the user can create a new analysis task according to the specific needs, and create a new task such as Figure 2 As shown, and save it to the database, the task management module can select the camera video within a specific time period for defect detection. The user can choose to start the analysis at a scheduled time or execute the analysis task immediately; when the analysis task arrives at the scheduled start time, the background will automatically transfer the camera video resources to the algorithm analysis module for defect analysis; the user can edit and delete the analysis task. When the analysis task is in progress, the user has the right to manually stop the analysis process; the algorithm analysis module is responsible for identifying defects; it supports frame skipping and speed analysis of videos that may remain static for a long time to improve analysis efficiency; the analysis results are reported to the alarm receiving and pushing module through the MQTT protocol; the ceiling alarm module first sends the defect The defect analysis data is stored in the database, and the processed defect information is then organized into an easy-to-read format and pushed to the user interface in real time via WebSockets technology. The system can dynamically generate groups based on tasks, cameras, and alarm content, allowing users to focus on observing the frequency and proportion of specific defects in the video. The video download module can download and save defect videos locally for subsequent maintenance and processing. The video download module also provides detailed video information, including the date, time, and defect type of the video, to facilitate user classification management and quick retrieval. Users can easily filter out the required video materials by timeline or defect type, greatly improving work efficiency.

[0034] The task management module of the substation defect video analysis system shown in the figure has analysis speeds of 2x, 4x, 8x, and 16x. The task management module can specify the types of defects to be analyzed, including but not limited to: crossing the line, not wearing a safety helmet, not wearing the required work clothes, smoking, not using a safety rope, people falling, oil stains on the ground, suspended matter in the air, bird nests, damaged or missing covers, fire smoke, water accumulation on the indoor floor, small animal intrusion, wall cracks, broken insulators, equipment icing, and broken silicone tubes.

[0035] The task management module of the substation defect video analysis system shown here has a function to prevent algorithm analysis service overload. Specifically, users can set an upper limit on the number of analysis tasks and the number of cameras to be analyzed simultaneously. When the number of analysis tasks reaches this limit, a new task will be placed in a waiting queue. Once the scheduled execution time arrives and the task queue conditions are met, the newly added task will begin the analysis process and enter the analyzing state. The upper limit of analysis tasks is taskLimit, the upper limit of camera analysis is cameraLimit, the number of tasks being analyzed is analyzeTask, the number of cameras being analyzed is analyzeCamera, the number of cameras included in the task currently being started is waitCamera, and whether the task is executable is canRun:

[0036] 1) When analyzeTask=0, it means there is no task being analyzed; then canRun=true;

[0037] 2) When analyzeTask>0; analyzeTask+1<=taskLimit, and waitCamera+analyseCamera<=cameraLimit; then canRun=true, otherwise canRun=true;

[0038] When canRun=true, it means that the task can be started, otherwise the task will re-enter the pending analysis state and wait for the next task to be started; whenever a task is completed or the user manually terminates a task, an attempt will be made to start the pending analysis task with the shortest execution time, entering the above judgment logic; during task analysis, the user can choose to manually terminate the task, making it enter the manual termination state, or wait for the task to complete naturally, thus entering the analysis completion state; for tasks that are not in the analysis state, the user can also choose to manually delete them, which will simultaneously delete the analysis results and related analyzed files; clicking the Details button allows the user to view the options set when the task was created to fully understand the specific content of the task.

[0039] The alarm time calculation of the algorithm analysis module of the substation defect video analysis system is shown as follows: let the alarm time be alarmTime, the video start time be videoStartTime, the number of frames in the video be frame, and the frame rate of the video be frameRate.

[0040]

[0041] By analyzing the number of frames where the defect occurs and the frame rate of the video, the specific time point (frame / frameRate) at which the defect occurs in the video is calculated. Based on this, the time of the defect alarm is the sum of the video start time and the duration of the defect location. This calculated alarm time helps locate the defect during subsequent video downloads. Users can adjust the download start and end times appropriately to download only the video clip containing the defect, quickly locating the problem without downloading the entire video file.

[0042] The analysis progress setting of the algorithm analysis module in the substation defect video analysis system shown is as follows: the analysis module regularly reports the video analysis progress of a single camera via the MQTT protocol every 5 seconds. Based on the tasks created by the user, which involve multiple cameras, the system integrates the video analysis progress of all cameras and displays a comprehensive analysis progress to the user. Let the overall progress be allProgress, the video analysis progress of the first camera is progress_1, the second is progress_2, and so on to the progress of the nth camera progress_n. In this way, the analysis progress of the entire task is calculated and updated and pushed to the user in real time. The formula is:

[0043]

[0044] By summarizing the analysis progress of each camera, dividing it by the total number of cameras, and finally multiplying it by 100, the overall percentage progress of the recording task can be calculated. This calculation method ensures that users can understand the progress of the entire recording analysis task in real time, so as to better arrange and manage work. At the same time, the system will also display this overall progress through a user-friendly interface, allowing users to intuitively see the completion status of the task in real time.

[0045] The alarm receiving and pushing module set up in the substation defect video analysis system shown has the function of repeated alarms caused by the persistence of defects; specifically: when MQTT receives the defect message analyzed by the algorithm, it will immediately store the original analysis results in the database, and push them to the user interface in real time through WebSockets technology for rendering and display, ensuring low latency and high reliability of data transmission; in order to prevent repeated alarms due to the persistence of defects, the system will query the corresponding task result group groupId based on the task identifier taskId, camera identifier cameraId and alarm content alarmContent in the analysis result; if groupId can be queried based on taskId, cameraId and alarmContent, the analysis result will be classified into the corresponding task group, and the last alarm time of the group and the last analysis time of the task will be updated to the current analysis result. The corresponding time, and the number of alarms is increased by 1; on the contrary, if the groupId cannot be queried, it means that the current analysis result has not been grouped. At this time, a new result group will be created, and the time of the most recent alarm and the time of the most recent analysis of the task will be recorded as the time corresponding to the current analysis result, and the number of alarms of the group will be initialized to 1; in the process of task group management, the system will also monitor the alarm frequency; if the number of alarms of a task group exceeds the set threshold within the preset time window, the system will automatically trigger a high-level alarm notification to remind the user to pay attention and deal with possible abnormal situations in time; in addition, the system will regularly analyze historical alarm data, identify common alarm patterns and potential risk points through machine learning algorithms, and provide data support for subsequent fault prevention and system optimization; through this series of management measures, the system can more effectively reduce the interference of invalid alarms, improve the work efficiency of operation and maintenance personnel, and ensure the stability and reliability of the entire monitoring system.

[0046] The video download module of the substation defect video analysis system shown in the figure is set to download video time as follows: alarm time is alarmTime, video download start time is videoDownLoadStartTime, and video download end time is videoDownLoadEndTime:

[0047] videoDownLoadStartTime=alarmTime-1

[0048] videoDownLoadEndTime=alarmTime+1

[0049] The default start and end times for video downloads are set to plus or minus 1 minute before and after the alarm time, simplifying the process of selecting the start and end times for video downloads. Users can preview and play the video and manually adjust the download start and end times to ensure that the video can be downloaded completely when the defect alarm occurs. This facilitates subsequent maintenance processing and allows users to use alarm videos for model training of analysis algorithms to further improve the accuracy of the analysis algorithms.

[0050] The above description is merely a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent variation based on the technical essence of the present invention shall still fall within the scope of protection claimed by the present invention.

Claims

1. A substation defect video analysis system, characterized by: The substation defect video analysis system is provided with a hard disk video recorder and camera ledger management module, a task management module, an algorithm analysis module, an alarm receiving and pushing module and a video download module; the hard disk video recorder and camera ledger management module provides the IP address of the hard disk video recorder and camera, the hard disk video recorder core driver type, the hard disk video recorder channel number to which the camera is connected, and the corresponding user name and password information, so as to generate the historical video RTSP address for the image algorithm analysis module to perform defect analysis on the video resources; the task management module can efficiently process analysis tasks, and users can create new analysis tasks according to specific needs and save them to the database. The task management module can select defect detection for camera videos within a specific time period, and users can choose to start analysis at leisure time or execute analysis tasks immediately; when the analysis task reaches the scheduled start time, the background will automatically transfer the camera video resources to the algorithm analysis module for defect analysis; users can edit and delete analysis tasks, and when the analysis task is in progress, users have the right to manually stop the analysis process; The algorithm analysis module is responsible for identifying defects; it supports frame-skipping and speed-up analysis of videos that may remain static for a long time to improve analysis efficiency; the analysis results are reported to the alarm receiving and pushing module via the MQTT protocol; The ceiling alarm module first stores the defect analysis data in the database, then organizes the processed defect information into an easy-to-read format and pushes it to the user interface in real time via WebSockets technology; The system can dynamically generate groups based on tasks, cameras, and alarm content, allowing users to focus on observing the frequency and proportion of specific defects in the video; the video download module can download and save defect videos locally to facilitate subsequent maintenance and processing; the video download module also provides detailed video information, including the date, time, and defect type of the video, to facilitate user classification management and quick retrieval; users can filter out the required video materials by timeline or defect type.

2. The substation defect video analysis system according to claim 1, characterized in that: The task management module of the substation defect video analysis system is configured with analysis speeds of 2x, 4x, 8x, and 16x. The task management module can specify the defect types to be analyzed, including but not limited to: crossing the line, not wearing a safety helmet, not wearing prescribed work clothes, smoking, not using a safety rope, people falling, oil stains on the ground, suspended matter in the air, bird nests, damaged or missing covers, fire smoke, water accumulation on the indoor floor, small animal intrusion, wall cracks, broken insulators, equipment icing, and broken silicone tubes.

3. The substation defect video analysis system according to claim 1, characterized in that: The task management module of the substation defect video analysis system has the function of avoiding overload of algorithm analysis service; specifically: Users can set the upper limit for the number of concurrent analysis tasks and the upper limit for the number of cameras being analyzed. When the number of analysis tasks reaches this limit, new tasks will be placed in a waiting queue. Once the scheduled execution time arrives and the task queue conditions are met, the newly added task will begin the analysis process and enter the analyzing state. The upper limit for analysis tasks is taskLimit, the upper limit for camera analysis is cameraLimit, the number of tasks being analyzed is analyzeTask, the number of cameras being analyzed is analyzeCamera, the number of cameras included in the task currently being started is waitCamera, and whether the task is executable is canRun: 1) When analyzeTask=0, it means there is no task being analyzed; then canRun=true; 2) When analyzeTask>0; analyzeTask+1<=taskLimit, and waitCamera+analyseCamera<=cameraLimit; then canRun=true, otherwise canRun=true; When canRun=true, it means that the task can be started, otherwise the task will re-enter the pending analysis state and wait for the next task to be started; whenever a task is completed or the user manually terminates a task, an attempt will be made to start the pending analysis task with the shortest execution time, entering the above judgment logic; during task analysis, the user can choose to manually terminate the task, making it enter the manual termination state, or wait for the task to complete naturally, thus entering the analysis completion state; for tasks that are not in the analysis state, the user can also choose to manually delete them, which will simultaneously delete the analysis results and related analyzed files; clicking the Details button allows the user to view the options set when the task was created to fully understand the specific content of the task.

4. The substation defect video analysis system according to claim 1, characterized in that: The alarm time calculation of the algorithm analysis module set in the substation defect video analysis system is as follows: let the alarm time be alarmTime, the video start time be videoStartTime, the number of frames in the video be frame, and the frame rate of the video be frameRate. By analyzing the number of frames where the defect occurs and the frame rate of the video, the specific time point (frame / frameRate) at which the defect occurs in the video is calculated. Based on this, the time of the defect alarm is the sum of the video start time and the duration of the defect location. This calculated alarm time helps locate the defect during subsequent video downloads. Users can adjust the download start and end times appropriately to download only the video clip containing the defect, quickly locating the problem without downloading the entire video file.

5. The substation defect video analysis system according to claim 1, characterized in that: The analysis progress of the algorithm analysis module set up in the substation defect video analysis system is specifically set as follows: the analysis module will regularly report the video analysis progress of a single camera through the MQTT protocol every 5 seconds; based on the task created by the user, involving multiple cameras, the system integrates the video analysis progress of all cameras and displays a comprehensive analysis progress to the user; let the overall progress be allProgress, the video analysis progress of the first camera is progress_1, the second is progress_2, and so on to the progress of the nth camera progress_n; in this way, the analysis progress of the entire task is calculated and updated and pushed to the user in real time; the formula is: By summarizing the analysis progress of each camera, dividing it by the total number of cameras, and finally multiplying it by 100, the overall percentage progress of the recording task can be calculated. This calculation method ensures that users can understand the progress of the entire recording analysis task in real time, so as to better arrange and manage work. At the same time, the system will also display this overall progress through a user-friendly interface, allowing users to intuitively see the completion status of the task in real time.

6. The substation defect video analysis system according to claim 1, characterized in that: The alarm receiving and pushing module set up in the substation defect video analysis system has the function of repeated alarms caused by the persistence of defects; specifically: when MQTT receives the defect message analyzed by the algorithm, it will immediately store the original analysis results in the database and push them to the user interface in real time through WebSockets technology for rendering and display, ensuring low latency and high reliability of data transmission; in order to prevent repeated alarms caused by the persistence of defects, the system will query the corresponding task result group groupId based on the task identifier taskId, camera identifier cameraId and alarm content alarmContent in the analysis result; if the groupId can be queried based on taskId, cameraId and alarmContent, the analysis result will be classified into the corresponding task group, and the latest alarm time and the latest analysis time of the task of the group will be updated to the corresponding time of the current analysis result, and the number of alarms will be increased by 1; on the contrary, if the groupId cannot be queried, it means that the current analysis result has not been grouped. At this time, a new result group will be created, and the latest alarm time and the latest analysis time of the task will be recorded as the time corresponding to the current analysis result, and the number of alarms of the group will be initialized to 1; During the task group management process, the system also monitors the alarm frequency; if the number of alarms for a task group exceeds the set threshold within the preset time window, the system will automatically trigger a high-level alarm notification, reminding users to pay attention and deal with possible abnormal situations in a timely manner; in addition, the system will regularly analyze historical alarm data, identify common alarm patterns and potential risk points through machine learning algorithms, and provide data support for subsequent fault prevention and system optimization; through this series of management measures, the system can more effectively reduce the interference of invalid alarms, improve the work efficiency of operation and maintenance personnel, and ensure the stability and reliability of the entire monitoring system.

7. The substation defect video analysis system according to claim 1, characterized in that: The video download module of the substation defect video analysis system is set to download video time as follows: the alarm time is alarmTime, the video download start time is videoDownLoadStartTime, and the video download end time is videoDownLoadEndTime: videoDownLoadStartTime=alarmTime-1 videoDownLoadEndTime=alarmTime+1 The default start and end times for video downloads are set to plus or minus 1 minute before and after the alarm time, simplifying the process of selecting the start and end times for video downloads. Users can preview and play the video and manually adjust the download start and end times to ensure that the video can be downloaded completely when the defect alarm occurs. This facilitates subsequent maintenance processing and allows users to use alarm videos for model training of analysis algorithms to further improve the accuracy of the analysis algorithms.

Images