Multi-platform integrated converter station operator and vehicle cooperative positioning method and system

Abstract

The application provides a multi-platform integrated converter station operator and vehicle cooperative positioning method and system, relates to the technical field of data processing, and comprises the following steps: constructing a converter station partition constraint graph; acquiring a converter station monitoring visualization model; performing cooperative operation accident suppression configuration on the converter station monitoring visualization model according to a personnel-vehicle work constraint layer; performing operation deviation accident suppression configuration on the converter station monitoring visualization model according to a personnel-vehicle irrelevant constraint layer; performing prohibition triggering accident suppression configuration on the converter station monitoring visualization model according to a personnel-vehicle prohibition constraint layer; and performing dynamic operation and maintenance of the converter station. The application solves the technical problem that, in the prior art, the positioning control and business management are disconnected, the static electronic fence is rigid, the human-vehicle cooperative decision is extensive, and the converter station operation and maintenance efficiency is affected, realizes human-vehicle hierarchical and graded accurate cooperative control and active and advanced intervention in accidents, and improves the converter station operation and maintenance efficiency.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to a method and system for collaborative positioning of converter station operators and vehicles that integrates multiple platforms. Background Technology

[0002] Digitalization is a key measure to improve power transmission efficiency and ensure the stable operation of the power system. As an important component of digital converter stations, the positioning system for outdoor personnel and special vehicles directly impacts the safety and efficiency of operations within the station. Currently, the positioning and control solutions for personnel and vehicles in converter stations generally combine multi-source positioning technology with static electronic fences. However, the positioning system has long been disconnected from business systems such as work permit management and access authorization. Positioning data is only output as isolated coordinate information, lacking deep integration and semantic analysis with business operations. This makes it difficult to dynamically identify the legality of personnel and vehicle behavior within specific areas based on real-time work tasks. Furthermore, electronic fences are typically set with fixed physical coordinates, failing to adapt flexibly to the effective and ineffective periods of work permits. This leads to frequent boundary crossing alarms by personnel with valid permits, resulting in numerous false alarms and weakening the effectiveness of safety control. Due to the lack of in-depth support from business semantics, existing human-vehicle collaborative decision-making often relies on a uniform distance threshold across the entire site for coarse alarms, failing to implement hierarchical differentiated judgment and risk suppression. This results in coarse human-vehicle collaborative decision-making, with a large number of inefficient or even invalid alarm messages flooding the monitoring platform. This not only seriously interferes with the continuity and focus of normal operation and maintenance, but also causes truly high-risk abnormal behaviors to be submerged in a massive number of false alarms, making it difficult to identify and handle them in a timely manner, posing a potential threat to the safe and efficient operation of the entire converter station.

[0003] In summary, existing technologies suffer from technical problems such as the disconnect between location control and business management, resulting in rigid static electronic fences and crude human-vehicle collaborative decision-making, which further affect the operation and maintenance efficiency of converter stations. Summary of the Invention

[0004] The purpose of this application is to provide a multi-platform integrated method and system for collaborative positioning of converter station personnel and vehicles, in order to solve the technical problems in the existing technology where the disconnect between positioning control and business management leads to rigid static electronic fences and crude human-vehicle collaborative decision-making, which further affects the operation and maintenance efficiency of converter stations.

[0005] To achieve the above objectives, this application provides a multi-platform integrated method and system for collaborative positioning of converter station operators and vehicles.

[0006] Firstly, this application provides a multi-platform integrated method for collaborative positioning of converter station operators and vehicles. This method is implemented through a multi-platform integrated system for collaborative positioning of converter station operators and vehicles. The method includes: acquiring access personnel and vehicle data, work permit data, positioning monitoring data, and video surveillance data of the converter station through multiple platforms; and constructing a converter station zoning constraint map based on the access personnel and vehicle data and the work permit data. The converter station zoning constraint map includes a personnel and vehicle work constraint layer, a personnel and vehicle unrelated constraint layer, and a personnel and vehicle... The system employs a vehicle prohibition constraint layer; performs dynamic calculation and fusion on the positioning monitoring data and the video surveillance data to obtain a converter station monitoring visualization model; configures collaborative operation accident suppression on the converter station monitoring visualization model according to the personnel and vehicle work constraint layer to obtain a first operation and maintenance strategy; configures operation deviation accident suppression on the converter station monitoring visualization model according to the personnel and vehicle irrelevant constraint layer to obtain a second operation and maintenance strategy; configures prohibited triggering accident suppression on the converter station monitoring visualization model according to the personnel and vehicle prohibition constraint layer to obtain a third operation and maintenance strategy; and combines the first operation and maintenance strategy and the second operation and maintenance strategy to perform dynamic operation and maintenance of the converter station.

[0007] Optionally, access control instructions are generated based on the access personnel and vehicle data, and dynamic access control management of the converter station is executed based on the access control instructions.

[0008] Optionally, the converter station is divided according to the work order data to determine work-related areas and work-independent areas; the work-independent areas are risk-classified to obtain work-independent primary areas and work-independent secondary areas; the work-related areas are time-divisionally mapped and bound according to the access personnel and vehicle data and the work order data to obtain the personnel and vehicle work constraint layer; the work-independent secondary areas are time-divisionally mapped and bound according to the access personnel and vehicle data and the work order data to obtain the personnel and vehicle unrelated constraint layer; the work-independent primary areas are time-divisionally mapped and bound according to the access personnel and vehicle data and the work order data to obtain the personnel and vehicle prohibition constraint layer; and the converter station zoning constraint map is generated by combining the personnel and vehicle work constraint layer and the personnel and vehicle unrelated constraint layer.

[0009] Optionally, occlusion blur detection is performed on the video surveillance data to obtain a video occlusion blur feature set; interpolation compensation is performed on the video surveillance data based on the video occlusion blur feature set to obtain optimized monitoring data; dynamic calculation is performed on the positioning monitoring data based on the optimized monitoring data to obtain positioning calculation results; and fusion modeling is performed based on the optimized monitoring data and the positioning calculation results to generate the converter station monitoring visualization model.

[0010] Optionally, the terminal positioning parameter K of the Kth target is read according to the positioning monitoring data, and the video positioning parameter K of the Kth target is read according to the optimized monitoring data, where K is a positive integer; the Kth positioning deviation is calculated based on the terminal positioning parameter K and the video positioning parameter K; it is determined whether the Kth positioning deviation is greater than or equal to the positioning deviation threshold; if the Kth positioning deviation is greater than or equal to the positioning deviation threshold, trajectory tracking is performed on the optimized monitoring data according to the video positioning parameter K to obtain the tracking result of the Kth target; the terminal positioning parameter K is correlated and corrected according to the tracking result of the Kth target to obtain the Kth positioning solution parameter, and the Kth positioning solution parameter is added to the positioning solution result.

[0011] Optionally, if the Kth positioning deviation is less than the positioning deviation threshold, the Kth terminal positioning parameter and the Kth video positioning parameter are fused together to output the Kth positioning solution parameter.

[0012] Optionally, the spatiotemporal mapping differential comparison of the converter station monitoring visualization model is performed based on the personnel and vehicle work constraint layer to generate a collaborative operation differential itemset; accident simulation is performed based on the collaborative operation differential itemset to construct a collaborative operation accident evolution chain with collaborative operation accident risk degree; a collaborative operation accident suppression decision set is configured based on the collaborative operation accident evolution chain; accident suppression simulation is performed on the converter station monitoring visualization model based on the collaborative operation accident suppression decision set to obtain simulation data for each suppression decision; based on the collaborative operation accident risk degree, accident risk reduction calculation is performed on the simulation data for each suppression decision to obtain each accident risk reduction degree; the collaborative operation accident suppression decision set is filtered based on each accident risk reduction degree to output the first operation and maintenance strategy.

[0013] Optionally, the personnel and vehicle-independent constraint layer is mapped to the converter station monitoring visualization model to identify explicit deviation triggering feature sets and implicit deviation triggering feature sets; the risk levels of the explicit deviation triggering feature sets and the implicit deviation triggering feature sets are evaluated respectively to obtain a deviation triggering risk level set; attention configuration fusion is performed on the explicit deviation triggering feature sets and the implicit deviation triggering feature sets based on the deviation triggering risk level set to obtain an explicit-implicit deviation triggering coupling graph; explicit deviation regression constraints are configured based on the explicit deviation triggering feature sets, and implicit deviation regression constraints are configured based on the implicit deviation triggering feature sets; accident suppression parameters are configured on the explicit-implicit deviation triggering coupling graph based on the explicit deviation regression constraints and the implicit deviation regression constraints to generate the second operation and maintenance strategy.

[0014] Optionally, the personnel and vehicle prohibition constraint layer is mapped to the converter station monitoring visualization model to identify explicit prohibition trigger feature sets and implicit prohibition trigger feature sets; based on the explicit prohibition trigger feature sets, explicit prohibition trigger accident paths are predicted, and based on the implicit prohibition trigger feature sets, implicit prohibition trigger accident paths are predicted; the explicit and implicit prohibition trigger accident paths are spatiotemporally correlated and fused to construct a prohibition trigger accident tree; and multi-node suppression parameters are configured on the prohibition trigger accident tree to generate the third operation and maintenance strategy.

[0015] Secondly, this application also provides a multi-platform integrated converter station personnel and vehicle collaborative positioning system for executing the multi-platform integrated converter station personnel and vehicle collaborative positioning method as described in the first aspect. The multi-platform integrated converter station personnel and vehicle collaborative positioning system includes: a data acquisition module for acquiring access personnel and vehicle data, work permit data, positioning monitoring data, and video surveillance data of the converter station through multiple platforms; a constraint graph construction module for constructing a converter station partition constraint graph based on the access personnel and vehicle data and the work permit data, the converter station partition constraint graph including a personnel and vehicle work constraint layer, a personnel and vehicle irrelevant constraint layer, and a personnel and vehicle prohibition constraint layer; and a dynamic solution and fusion module for processing the positioning monitoring data. The data and the video surveillance data are dynamically calculated and fused to obtain a converter station monitoring visualization model; a collaborative operation accident suppression configuration module is used to configure collaborative operation accident suppression on the converter station monitoring visualization model according to the personnel and vehicle work constraint layer to obtain a first operation and maintenance strategy; an operation deviation accident suppression configuration module is used to configure operation deviation accident suppression on the converter station monitoring visualization model according to the personnel and vehicle irrelevant constraint layer to obtain a second operation and maintenance strategy; a prohibition triggering accident suppression configuration module is used to configure prohibition triggering accident suppression on the converter station monitoring visualization model according to the personnel and vehicle prohibition constraint layer to obtain a third operation and maintenance strategy, and the dynamic operation and maintenance of the converter station is executed in combination with the first operation and maintenance strategy and the second operation and maintenance strategy.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages: by constructing a three-layer partition constraint map driven by access data and work tickets, and deeply coupling it with a visualization model that integrates video-positioning dynamic calculation, multi-level intelligent decision-making is configured for the work constraint layer, irrelevant constraint layer and prohibited constraint layer respectively to achieve precise collaborative control of personnel and vehicles at different levels and proactive intervention in accidents, thereby improving the operation and maintenance efficiency of converter stations.

[0017] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the multi-platform integrated converter station operator and vehicle collaborative positioning method of this application.

[0020] Figure 2 This is a curve showing the dynamic evolution and strategy reduction of converter station operation risks in the multi-platform integrated converter station operator and vehicle collaborative positioning method of this application.

[0021] Figure 3 This is a schematic diagram of the multi-platform integrated converter station operator and vehicle collaborative positioning system of this application.

[0022] Figure labeling: Data acquisition module 11, constraint graph construction module 12, dynamic solution fusion module 13, collaborative operation accident suppression configuration module 14, operation deviation accident suppression configuration module 15, and accident suppression configuration module to prevent triggering 16. Detailed Implementation

[0023] This application addresses the technical problem in existing technologies where the disconnect between location control and business management leads to rigid static electronic fences and inefficient human-vehicle collaborative decision-making, further impacting the operation and maintenance efficiency of converter stations. It provides a multi-platform integrated method and system for collaborative positioning of personnel and vehicles at converter stations. By constructing a three-layer partitioned constraint graph driven by access data and work orders, and deeply coupling it with a visualization model that integrates video and location dynamic calculations, multi-level intelligent decision-making is configured for the work constraint layer, irrelevant constraint layer, and prohibited constraint layer, respectively. This enables precise, hierarchical collaborative control of personnel and vehicles, as well as proactive, advance intervention in accidents, thereby improving the operation and maintenance efficiency of converter stations.

[0024] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0025] Example 1, please refer to the appendix. Figure 1 This application provides a multi-platform integrated method for the collaborative positioning of converter station operators and vehicles. The method is applied to a multi-platform integrated system for the collaborative positioning of converter station operators and vehicles. Specifically, the method includes the following steps: Data on personnel and vehicles entering the converter station, work permits, location monitoring data, and video surveillance data are obtained through multiple platforms.

[0026] Furthermore, this application also includes the following steps: generating access control instructions based on the access personnel and vehicle data, and executing dynamic access control management of the converter station based on the access control instructions.

[0027] Specifically, a multi-source data acquisition and aggregation platform is deployed at the converter station, connecting to the access control system, work permit management system, positioning base stations and terminal data backhaul channels, and the streaming media server of the video surveillance system via standardized interfaces. The data aggregation platform uses a unified time base and spatial coordinate system to perform timestamp alignment and coordinate normalization on all data streams. Regarding access personnel and vehicle data, registration information for all currently valid authorized personnel and vehicles is retrieved from the access control system on a regular schedule, and primary key indexes are created based on personnel and vehicle numbers. For work permit data, the system proactively queries the work permit management system for currently effective permits and permits that will become effective within a preset future time period, extracting the task type, permitted area boundary coordinates, permitted personnel list, and permitted vehicle list for each permit. Simultaneously, it monitors permit status change events to achieve real-time data updates. Regarding positioning monitoring data, multiple positioning data receiving gateways deployed in the converter station area receive raw positioning data reported at fixed frequencies by personnel-worn positioning tags and vehicle-mounted positioning terminals via wireless communication. This data includes satellite observations, pseudorange, carrier phase, UWB ranging values, and IMU angular velocity and acceleration. After preliminary preprocessing at the edge, the data is transmitted back to the aggregation platform in the form of data packets. Regarding video surveillance data, real-time RTSP video streams from each camera are retrieved through a video stream access gateway, and cached and cataloged according to each frame and its corresponding timestamp.

[0028] Detailed records of all currently authorized personnel and vehicles are extracted from the access control data aggregated from multiple platforms. For each record, the effective start and end times of access, the set of allowed area codes, and the corresponding access control point number are parsed. Simultaneously, equipment ledger information for all access control points in the converter station is retrieved, including the number of each access control gate or barrier, its installation location coordinates, the name of the controlled physical entrance, and the corresponding internal area code.

[0029] Based on the current time of the global clock, a time validity check is performed on each access record. If the current time falls within the interval between the record's valid start and end times, the record is marked as a currently valid access record; otherwise, it is marked as a currently invalid access record. All currently valid access records are grouped and aggregated according to the access control point number. For each access control point, a list of all personnel and vehicles authorized to enter its controlled area through that access control point during the current time period is compiled, forming an authorized access whitelist for that access control point.

[0030] For each access control point, a corresponding access control command is generated. For access control points with an authorized access whitelist that is not empty, a verification-opening command is generated, requiring the access control point to enter authentication mode. When a person or vehicle arrives at the access control point and presents their credentials, the access controller compares the identity identifier in the credentials with the locally cached authorized access whitelist. If a match is found, the door is opened and access is logged; otherwise, the access point remains locked and an authentication failure alarm is triggered. For access control points with an empty authorized access whitelist (meaning no valid access record has been granted access during the current time period), a forced locking command is generated, requiring the access control point to be completely locked for the duration of the command. No access is permitted regardless of credentials, and the access control point enters a restricted state, automatically triggering an unauthorized entry alarm when a person or vehicle attempts to pass.

[0031] All generated access control commands are transmitted to the corresponding access controller terminals via the wireless communication network. Upon receiving a command, each access controller switches to the corresponding operating mode according to the command type and continues to execute that mode until the next update command is issued. Simultaneously, a continuous refresh mechanism for access control commands is established, re-executing the entire process at preset time intervals to ensure that when access records change, such as new personnel gaining access authorization or existing authorizations expiring, the access control device's control policy can be dynamically updated with minimal delay.

[0032] During dynamic access control management, records of access events reported by each access control point are continuously collected, including the time of each access opening, the identity of the person passing through, the type of credentials held, the verification result, and the time and identity information of each attempt to refuse access. This data is used to cross-verify with location monitoring data to confirm that personnel and vehicles entering the station do indeed have access control records, avoiding inconsistencies where location data shows the person is in the station but access control has no entry record.

[0033] Based on the access personnel and vehicle data and the work ticket data, a converter station zoning constraint map is constructed. The converter station zoning constraint map includes a personnel and vehicle work constraint layer, a personnel and vehicle irrelevant constraint layer, and a personnel and vehicle prohibition constraint layer.

[0034] Furthermore, this application also includes the following steps: dividing the converter station according to the work order data to determine work-related areas and work-independent areas; classifying the work-independent areas by risk to obtain work-independent primary areas and work-independent secondary areas; performing time-sharing mapping and binding of the work-related areas according to the access personnel and vehicle data and the work order data to obtain the personnel and vehicle work constraint layer; performing time-sharing mapping and binding of the work-independent secondary areas according to the access personnel and vehicle data and the work order data to obtain the personnel and vehicle unrelated constraint layer; performing time-sharing mapping and binding of the work-independent primary areas according to the access personnel and vehicle data and the work order data to obtain the personnel and vehicle prohibition constraint layer; and generating the converter station partition constraint map by combining the personnel and vehicle work constraint layer and the personnel and vehicle unrelated constraint layer.

[0035] Specifically, the system reads all active work order data at the current moment, parses each work order to extract its work area boundary description information. Each work order contains a textual or graphical description of the physical space occupied by the task. Through natural language parsing or spatial data reading, this boundary description information is converted into a closed polygon coordinate sequence in a unified coordinate system within the station. Taking the entire physical space enclosed by the converter station's perimeter wall as the complete set, a set difference operation is performed between this complete set and the polygons of all work order work areas. The result of the set difference operation is determined as a work-independent region, while the union of all work order work area polygons is determined as a work-related region. The work-related area is the geographic space area explicitly specified in the currently effective work order for performing a specific work task. It is the physical carrier of the legal work activities permitted by the work order. All location coordinates within the area are recorded as closed polygons or composite area boundaries composed of multiple sub-polygons. The work-irrelevant area is the entire physical space area of ​​the converter station, excluding all work areas specified in currently effective work orders. There are no currently effective work tasks within the work-irrelevant area, and personnel or vehicles entering this area do not have the legality of work granted by the work order.

[0036] Risk classification is implemented for work-irrelevant areas. A pre-set hazard level map of equipment within the station is used to classify all equipment areas in the station into basic hazard levels based on their operating status and physical attributes. For each spatial point within a work-irrelevant area, the hazard level attribute of its corresponding equipment area is queried and categorized according to the preset hazard level mapping rules. Spatial locations belonging to equipment in energized operation with a voltage level of 220kV or above, high-voltage bushing areas of converter transformers, gas chamber areas of 500kV gas-insulated fully enclosed combined electrical equipment, and high-voltage capacitor areas of DC filter banks are all classified as high-risk locations. Spatial locations belonging to station inspection passages, paved roads, non-equipment installation areas, areas surrounding out-of-use equipment, green belts, or office auxiliary areas are all classified as low-risk locations. The spatial set of high-risk locations is marked as work-irrelevant Level 1 areas, and the spatial set of low-risk locations is marked as work-irrelevant Level 2 areas. Work-irrelevant Level 1 and Level 2 areas constitute a complete division of work-irrelevant areas, with no overlap and complementing each other to form a complete set.

[0037] All work order operation areas within the work-related region are associated and bound to the corresponding list of authorized personnel and vehicles for each work order. The start and end times of each work order's validity are then added as time-dimensional constraints to this binding relationship, generating a personnel and vehicle work constraint layer. A universal access binding relationship is established between the work-independent secondary regions and all personnel and vehicles holding valid station access qualifications. The validity period of the personnel or vehicle's access authorization is used as a time-dimensional constraint, generating a personnel and vehicle unrelated constraint layer. The work-independent primary regions are designated as prohibited areas, and an unconditional prohibition mapping is established between these areas and all personnel and vehicles; that is, no personnel or vehicle is authorized to enter these areas. This layer remains valid throughout the time dimension without an end time, generating a personnel and vehicle prohibition constraint layer.

[0038] The three layers of spatial data—personnel and vehicle operational constraint layer, personnel and vehicle irrelevant constraint layer, and personnel and vehicle prohibited constraint layer—along with their corresponding personnel and vehicle mapping relationships, are overlaid and stored in a unified spatial data structure to form a converter station zoning constraint map. Each constraint layer contains complete attribute fields such as spatial boundary coordinate sequence, authorized personnel list, authorized vehicle list, start valid time, and end valid time. Each layer is assigned a unique identifier number, and the three layers of data are combined and packaged to form a complete converter station zoning constraint map data file.

[0039] Dynamic calculation and fusion are performed on the location monitoring data and the video surveillance data to obtain a converter station monitoring visualization model.

[0040] Furthermore, this application also includes the following steps: performing occlusion blur detection on the video surveillance data to obtain a video occlusion blur feature set; performing interpolation compensation on the video surveillance data based on the video occlusion blur feature set to obtain optimized monitoring data; performing dynamic calculation on the positioning monitoring data based on the optimized monitoring data to obtain positioning calculation results; and performing fusion modeling based on the optimized monitoring data and the positioning calculation results to generate the converter station monitoring visualization model.

[0041] Furthermore, this application also includes the following steps: reading the terminal positioning K-th parameter of the K-th target based on the positioning monitoring data, and reading the video positioning K-th parameter of the K-th target based on the optimized monitoring data, where K is a positive integer; calculating the K-th positioning deviation based on the terminal positioning K-th parameter and the video positioning K-th parameter; determining whether the K-th positioning deviation is greater than or equal to a positioning deviation threshold; if the K-th positioning deviation is greater than or equal to the positioning deviation threshold, performing trajectory tracking on the optimized monitoring data based on the video positioning K-th parameter to obtain the tracking result of the K-th target; performing correlation correction on the terminal positioning K-th parameter based on the tracking result of the K-th target to obtain the K-th positioning solution parameter, and adding the K-th positioning solution parameter to the positioning solution result.

[0042] Furthermore, this application also includes the following steps: if the Kth positioning deviation is less than the positioning deviation threshold, the Kth terminal positioning parameter and the Kth video positioning parameter are fused together, and the Kth positioning solution parameter is output.

[0043] Specifically, the system acquires the current video surveillance data frame sequence and performs occlusion blur detection on each frame. Each frame is divided into multiple detection blocks, and image sharpness evaluation metrics are calculated for each block, including but not limited to spatial domain sharpness values ​​based on grayscale variance, frequency domain high-frequency component proportions based on discrete cosine transform, and texture intensity values ​​based on local contrast. The sharpness metrics of each block are compared with a preset normal sharpness benchmark. When the sharpness metric of a block is lower than a set proportion of the benchmark, that block is marked as an occluded blur region. These occluded blur regions undergo connected component merging, grouping spatially adjacent regions with similar blur levels into a single occlusion blur feature. The position and range of this feature in the image coordinate system are recorded, and simultaneously mapped to the station's spatial coordinate system according to the camera calibration parameters, forming a complete video occlusion blur feature set. For example, at 8:32:00:250, occlusion blur detection is performed on the current frame image of camera CAM07, covering the northwest corner of area B. The image frame was divided into 48 detection blocks, each 120 pixels × 120 pixels. The block group located in the upper right corner of the image, with pixel coordinates ranging from 960 to 1200 horizontally and 480 to 720 vertically, corresponds to approximately 175.0m to 185.0m and 352.0m to 360.0m within the camera's internal spatial coordinates. The grayscale variance obtained from the sharpness detection was 42.8, while the normal sharpness benchmark for this camera is 156.3. 42.8 is less than 50% of the benchmark, and a total of 6 blocks were marked as occluded blurry areas. The internal spatial coordinate range of 175.0m to 185.0m and 352.0m to 360.0m within these areas was recorded as video occlusion blurry features and added to the feature set.

[0044] For regions identified as occluded or blurred, the system identifies the targets expected to exist within these regions. Using the target's position sequence across several consecutive frames prior to the occlusion, linear extrapolation is used to predict the target's position in the current frame. The target's reported terminal position in the location monitoring data is then used as an auxiliary reference. A weighted average is applied between the motion extrapolation-predicted position and the terminal-reported position, with the weighting coefficient dynamically adjusted based on the severity of the occlusion blur—the more severe the blur, the greater the weight given to the terminal's location. After interpolation compensation, the target position within the occluded region is fully filled in. This supplemented data is then merged with the normal video detection data from the unoccluded region to form optimized monitoring data.

[0045] Each target within the converter station's monitoring range is numbered and labeled as target K, with K starting from 1 and sequentially increasing to the total number of monitored targets N. For target K, the terminal positioning parameter K for the target at the current moment is read from the positioning monitoring data, and the video positioning parameter K for the same target at the same moment is read from the optimized monitoring data. Based on the coordinate components in these two parameters, the Kth positioning deviation is calculated, which is the Euclidean distance between the terminal positioning coordinates and the video positioning coordinates. If either the terminal positioning parameter K or the video positioning parameter K is missing or invalid, the fusion processing for that target is skipped, and the available single-source data is directly used as the positioning solution parameter for that target.

[0046] The system determines whether the Kth positioning deviation is greater than or equal to the positioning deviation threshold. If it is, the terminal positioning data is deemed to have a significant error in the current environment, and the video positioning coordinates are used as a reliable benchmark. Using the video positioning coordinates at that moment as the starting reference position, the system traces the target's position coordinate sequence back through several recent consecutive frames in the optimized monitoring data. A linear fit is performed on this sequence using the least squares method to obtain the target's velocity and direction vectors within the most recent time window, and the fitting result is recorded as the Kth target tracking result. Based on the velocity and direction vectors in this tracking result, the Kth parameter of the terminal positioning is correlated and corrected: a correction value is superimposed on the terminal positioning coordinates before correction. This correction value has the same direction as the velocity vector and its magnitude is a weighted combination of the projection component of the current terminal positioning coordinates relative to the video positioning coordinates in the velocity vector direction and the average deviation of the recent frames. The corrected coordinates are then used as the Kth positioning solution parameter.

[0047] If the Kth positioning deviation is less than the positioning deviation threshold, the terminal positioning data is deemed to have high reliability in the current environment, and the video positioning data also provides reliable visual verification. The terminal positioning coordinates and the video positioning coordinates are weighted and averaged according to preset weights. The preset weights are dynamically determined based on the horizontal accuracy factor of the terminal positioning and the confidence score of the video detection. Finally, the weighted average coordinates are output as the Kth positioning solution parameter.

[0048] After generating the localization parameters for each target, these parameters, along with the target number, timestamp, and processing mode flag, are added to the localization result dataset. Once all N targets have undergone the above processing, the complete localization result is output. The localization result is a complete location information dataset formed by calculating the localization parameters for each identified target and then summarizing and arranging the Kth localization parameter of each target according to its target number.

[0049] The 3D building information model and electrical equipment layout model of the converter station are loaded, containing the precise spatial location and 3D geometric dimensions of all static facilities such as buildings, equipment foundations, supports, busbars, and pipe corridors. The static model serves as the spatial base map for the visualization model. A zone constraint map of the converter station is overlaid on this base map, with semi-transparent color blocks indicating the spatial extent of the personnel / vehicle working constraint layer, the irrelevant constraint layer, and the prohibited constraint layer, each layer distinguished by a different color. Each target from the positioning calculation results is plotted as an icon at its corresponding coordinate position in the model. The shape and color of each target's icon are differentiated according to its category and the type of the current constraint layer; for example, personnel are represented by circular icons, vehicles by square icons, targets within the working layer are displayed in green, targets within the irrelevant layer are displayed in yellow, and targets approaching the prohibited layer are displayed in flashing red.

[0050] Each target icon is labeled with its target number, personnel name or vehicle license plate number, current speed, etc., and its historical trajectory line for the most recent few seconds is plotted. This model data is packaged in a standard format and pushed to the dispatch center's large-screen display terminal and the 3D digital twin platform, continuously updated at a refresh rate of no less than ten times per second, forming a dynamic converter station monitoring visualization model. The converter station monitoring visualization model is the final output of the fusion modeling steps, presented on the dispatch center's large screen or the 3D digital twin platform as a real-time digital mirror of the converter station's status. It is not merely a simple overlay of static scenes and dynamic target locations, but also includes each target's identity information, authorization status, movement trend, and spatial topological relationship with static equipment.

[0051] Based on the personnel and vehicle work constraint layer, the converter station monitoring visualization model is configured with collaborative operation accident suppression to obtain the first operation and maintenance strategy.

[0052] Furthermore, this application also includes the following steps: performing spatiotemporal mapping differential comparison on the converter station monitoring visualization model based on the personnel and vehicle work constraint layer to generate a collaborative operation differential itemset; performing accident simulation based on the collaborative operation differential itemset to construct a collaborative operation accident evolution chain with collaborative operation accident risk degree; configuring a collaborative operation accident suppression decision set based on the collaborative operation accident evolution chain; performing accident suppression simulation on the converter station monitoring visualization model based on the collaborative operation accident suppression decision set to obtain simulation data for each suppression decision; calculating accident risk reduction based on the collaborative operation accident risk degree to obtain each accident risk reduction degree; filtering the collaborative operation accident suppression decision set based on each accident risk reduction degree, and outputting the first operation and maintenance strategy.

[0053] Specifically, the system extracts real-time motion data of all targets and vehicles currently within the personnel and vehicle work constraint layer, including the current coordinates, instantaneous velocity vector, heading angle, and acceleration vector of each target. Simultaneously, it reads the work order data corresponding to this work constraint layer, extracting the standard operating procedure (SOP) specified in the SOP. This SOP details the nominal position of each worker and each vehicle during the operation, the movement path they must follow, and the maximum permissible speed and minimum safety distance.

[0054] After aligning the actual state with the standard scheme using a unified time reference, a spatiotemporal mapping differential comparison is performed on each target within the working constraint layer at each time step. For each sampling time, the position deviation vector between the target's actual coordinates and the coordinates specified in the standard scheme is calculated, and the modulus is taken to obtain the position deviation. The difference between the actual velocity magnitude and the standard velocity magnitude is calculated to obtain the velocity deviation. The angle between the actual velocity direction and the standard velocity direction is calculated to obtain the direction deviation. The position deviation, velocity deviation, and direction deviation are compared with preset position deviation tolerance thresholds, velocity deviation tolerance thresholds, and direction deviation tolerance thresholds, respectively. When any deviation exceeds the corresponding tolerance threshold, the target is recorded as a differential item at that time. The target number, deviation occurrence time, deviation occurrence location, position deviation, velocity deviation, direction deviation, and a comprehensive deviation severity score are written into the differential item record. After traversing all sampling times for all targets within the working constraint layer, all differential item records are summarized to form a collaborative operation differential item set.

[0055] For each differential item in the collaborative operation differential item set, an accident simulation is performed. Starting with the current time recorded in the differential item and its current position, and using the target's current velocity and acceleration vectors as initial motion conditions, a uniform acceleration motion model is used for time extrapolation. The simulation time window is preset to 10 to 30 seconds. During the simulation, the motion states of all other targets within the work constraint layer are considered simultaneously. For each pair of future times, the minimum spatial distance between any two targets is calculated. When it is detected that the minimum distance between a pair of targets will be less than a preset safe distance threshold at a certain moment within the simulation time window, the complete time sequence and position sequence from the initial differential item to the occurrence of this minimum distance event are recorded, forming a collaborative operation accident evolution chain. The collaborative operation accident risk degree is calculated for each evolution chain by multiplying the conditional risk probabilities of each intermediate node from the starting node to the final node. The conditional risk probability of each node is determined by combining the ratio of the distance between targets at that node to the safe distance threshold and the relative velocity of the targets at that moment. Assign a calculated risk value to each evolutionary chain.

[0056] Based on the risk type and location in the collaborative operation accident evolution chain, a corresponding collaborative operation accident suppression decision set is configured. For evolution chains where the risk of rear-end collision is caused by excessive target speed, a deceleration suppression decision is configured; for evolution chains where path intersection is caused by target heading deviation, a yaw correction suppression decision is configured; and for evolution chains where congestion risk is caused by multiple targets converging in a narrow space, a timing adjustment suppression decision is configured. These various decisions are then organized into a candidate decision set according to a structured format of target, triggering condition, and execution parameters.

[0057] Each candidate decision in the collaborative operation accident suppression decision set is input into the accident suppression simulation engine. The simulation engine uses the current state of the converter station monitoring visualization model as initial conditions, applies the control actions specified by the candidate decision to be simulated, and drives the model to evolve forward according to kinematic laws over the complete time window. The simulation engine records the simulated trajectories of each target after the decision is applied, the minimum distance sequence between any two targets in the simulated trajectory, and the final determination of whether a collision event occurs. The simulation results corresponding to each candidate decision are compiled into suppression decision simulation data.

[0058] Based on the original risk level of each collaborative operation accident evolution chain and the residual risk level measured after simulation of each candidate decision, the accident risk reduction degree of each candidate decision is calculated, which is the original risk level minus the residual risk level. Simultaneously, a decision side effect coefficient is introduced into the risk reduction degree calculation to quantify the adverse impact of executing the decision on normal operation efficiency. The final net risk reduction degree is the original risk reduction degree minus the side effect coefficient multiplied by a preset penalty weight. After calculating the net risk reduction degree of each candidate decision, all candidate decisions are sorted in descending order. Decisions are selected sequentially from the sorted results, and the total net risk reduction degree of the selected decision set is calculated. Selection stops when the marginal reduction gain from adding the next decision is less than a preset minimum effective gain threshold, and the selected decision set is taken as the final decision combination. The specific execution parameters of all decisions in this decision combination are encapsulated in a unified instruction format and output as the first operation and maintenance strategy. The first operation and maintenance strategy contains specific collaborative safety control instructions for each target and vehicle within the work constraint layer.

[0059] Based on the personnel and vehicle-independent constraint layer, the converter station monitoring visualization model is configured to suppress operational deviation accidents, and a second operation and maintenance strategy is obtained.

[0060] Furthermore, this application also includes the following steps: mapping the personnel and vehicle-independent constraint layer to the converter station monitoring visualization model, identifying explicit deviation triggering feature sets and implicit deviation triggering feature sets; evaluating the risk level of the explicit deviation triggering feature sets and the implicit deviation triggering feature sets respectively, obtaining a deviation triggering risk level set; performing attention configuration fusion on the explicit deviation triggering feature sets and the implicit deviation triggering feature sets according to the deviation triggering risk level set, obtaining an explicit-implicit deviation triggering coupling graph; configuring explicit deviation regression constraints based on the explicit deviation triggering feature sets, configuring implicit deviation regression constraints based on the implicit deviation triggering feature sets; configuring accident suppression parameters on the explicit-implicit deviation triggering coupling graph based on the explicit deviation regression constraints and the implicit deviation regression constraints, generating the second operation and maintenance strategy.

[0061] Specifically, the spatial boundary coordinate data of the personnel and vehicle-independent constraint layers are read and overlaid onto the base map of the converter station monitoring visualization model. This allows the position coordinates of each dynamic target in the visualization model to be synchronously queried relative to the boundary of the independent constraint layer during updates. The real-time spatial determination of each target's position in each frame of the visualization model is performed. For each target, the vertical distance from its current position coordinates to the common boundary between the nearest independent constraint layer and the prohibited constraint layer, as well as the vertical distance to the common boundary between the nearest independent constraint layer and the working constraint layer, are calculated.

[0062] Continuously monitor whether the position coordinates of each target have crossed any common boundary line and entered the prohibited constraint layer or its unauthorized working constraint layer. Once a boundary crossing event is detected, immediately record the boundary crossing target number, the time of boundary crossing, the boundary crossing position coordinates, the instantaneous velocity vector at the time of boundary crossing, the angle between the boundary crossing direction and the boundary normal, and the type of area entered after boundary crossing. Combine these into an explicit deviation trigger feature record and store it in the explicit deviation trigger feature set.

[0063] For targets that have not yet crossed the boundary, a multi-criteria detection is performed. The first criterion is the direction criterion, which calculates the angle between the target's current velocity direction and the normal direction of the nearest prohibited constraint layer boundary. If the angle is less than a preset angle threshold and the target velocity is greater than a preset silent velocity threshold, a direction warning is triggered. The second criterion is the trend criterion, which calculates the target's average movement direction over the most recent consecutive frames. If the standard deviation of the average direction is less than a preset stability threshold and the average direction continues to point towards the prohibited area, a trend warning is triggered. The third criterion is the distance criterion, which calculates the target's current distance from the nearest prohibited area boundary. If the distance is decreasing at a rate exceeding a preset rate threshold and the current distance is less than a preset long-distance warning threshold, a distance warning is triggered. When at least two of the three criteria are met simultaneously, the target is marked as a latent deviation trigger feature, and the target number, current coordinates, current velocity vector, distance from the prohibited area boundary, predicted estimated boundary crossing time, and predicted boundary crossing point location are recorded and stored in the latent deviation trigger feature set.

[0064] Risk assessments were performed on both the explicit and implicit deviation trigger feature sets. For explicit deviation trigger features, the risk score was calculated as follows: normalized cross-boundary speed × speed weight + entry area hazard level coefficient × area weight + normalized duration of stay after cross-boundary × duration weight. These three values ​​were weighted, summed, and normalized to a range of 0 to 1, where the sum of the three weights was 1. For implicit deviation trigger features, the risk score was calculated as follows: normalized reciprocal of the current distance from the restricted area boundary × distance weight + normalized current approach speed × speed weight + directional stability score × directional weight. These three values ​​were weighted, summed, and normalized to a range of 0 to 1, where distance weight + speed weight + directional weight = 1. The risk scores for all features were then aggregated by feature number to form a deviation trigger risk level set.

[0065] Attention configuration fusion is performed based on the deviation trigger risk level set. Explicit and implicit deviation trigger features are treated as two types of nodes, constructing a bipartite graph structure. For each explicit deviation feature node, its spatial distance and time interval with all implicit deviation feature nodes are calculated. If the spatial distance is less than a preset spatial association radius and the time interval is less than a preset temporal association window, an undirected edge is established between the two nodes. Using the risk level score of each node as the query vector and key vector input in the attention mechanism, the attention weight coefficient of each node relative to its neighboring nodes is calculated through scaled dot product attention, reflecting the influence of the node on its neighbors in the coupled graph. The calculated attention weight coefficients are assigned to each node and each edge, forming a complete weighted explicit-implicit deviation trigger coupled graph. The core function of the explicit-implicit deviation trigger coupled graph is to place currently occurring boundary violations and potential boundary violations approaching the restricted area within the same computational framework for correlation analysis, revealing the interaction between the two.

[0066] For each explicit deviation trigger feature, corresponding regression constraint parameters are set according to the type of area it entered. If it entered a prohibited constraint layer, the most stringent regression constraint is configured, immediately triggering a continuous audible and visual alarm until the target completely exits the restricted area and the exit time exceeds the preset confirmation duration. Simultaneously, a safety event record is automatically generated and pushed to the dispatch center. If it illegally entered a working constraint layer, a medium-strength regression constraint is configured, triggering a voice prompt to inform the target that it has crossed the boundary and instructing it to return to the unrelated constraint layer along the shortest path. Implicit deviation regression constraints are configured based on the implicit deviation trigger feature set. For each implicit deviation trigger feature, hierarchical and progressive regression constraint parameters are configured according to its current distance from the restricted area boundary. When the distance is greater than the first warning distance, a level one constraint is configured, displaying only a directional arrow on the terminal. When the distance is less than or equal to the first warning distance but greater than the second warning distance, a level two constraint is configured, playing a voice warning and adding a vibration alert. When the distance is less than or equal to the second warning distance, a level three constraint is configured, triggering strong vibration and continuous voice warnings and automatically notifying the person in charge of the nearby area.

[0067] The configured explicit and implicit deviation regression constraints are applied to the explicit-implicit deviation triggering coupling graph. Each node in the graph is assigned a corresponding suppression parameter value based on its risk level score and node type. Nodes with higher risk levels have longer warning trigger distances, louder warning volumes, and more forceful regression actions. The suppression parameters for all nodes are then aggregated to form a complete second operation and maintenance strategy, which is then output. This second strategy includes the specific execution parameters for explicit and implicit deviation regression constraints, as well as a configuration scheme for accident suppression parameters based on the coupling graph. It serves as an output instruction set for hierarchical control of deviation behavior within the irrelevant constraint layer.

[0068] Based on the personnel and vehicle prohibition constraint layer, the monitoring and visualization model of the converter station is configured to prevent accident triggering, and a third operation and maintenance strategy is obtained. The dynamic operation and maintenance of the converter station is then executed in combination with the first operation and maintenance strategy and the second operation and maintenance strategy.

[0069] Furthermore, this application also includes the following steps: mapping the personnel and vehicle prohibition constraint layer to the converter station monitoring visualization model, identifying explicit prohibition trigger feature sets and implicit prohibition trigger feature sets; predicting explicit prohibition trigger accident paths based on the explicit prohibition trigger feature sets, and predicting implicit prohibition trigger accident paths based on the implicit prohibition trigger feature sets; performing spatiotemporal correlation fusion on the explicit prohibition trigger accident paths and the implicit prohibition trigger accident paths to construct a prohibition trigger accident tree; configuring multi-node suppression parameters on the prohibition trigger accident tree to generate the third operation and maintenance strategy.

[0070] Specifically, the personnel and vehicle prohibition constraint layer is mapped onto the converter station monitoring visualization model. Specifically, three concentric early warning zones are automatically generated around the boundary of the prohibition constraint layer according to a preset zone radius. The first-level zone is a ring-shaped area extending from a certain distance from the boundary line to another relatively close distance; the second-level zone is a ring-shaped area extending from a relatively close distance to an even closer distance from the boundary line; and the third-level zone is a ring-shaped area extending from an even closer distance to the boundary line itself. The boundary data of the prohibition constraint layer boundary and the third-level early warning zones are all overlaid onto the converter station monitoring visualization model.

[0071] For each dynamic target in the converter station monitoring visualization model, the vertical distance from its current position coordinates to the nearest prohibited constraint layer boundary line is calculated, and the target's current defense zone is determined based on this distance value. If the vertical distance is greater than the outer boundary distance of the first-level defense zone, the target is determined to be in a safe area; if the vertical distance is between the outer boundary distance and the inner boundary distance of the first-level defense zone, the target is determined to be in the first-level defense zone; if the vertical distance is between the outer boundary distance and the inner boundary distance of the second-level defense zone, the target is determined to be in the second-level defense zone; if the vertical distance is between the outer boundary distance and the inner boundary distance of the third-level defense zone, the target is determined to be in the third-level defense zone; if the vertical distance is negative, meaning the target has crossed the boundary line, the target is determined to be in a restricted area.

[0072] For targets identified as being within the restricted area, record the target's number, the precise time it crossed the boundary line, the coordinates of the crossing point, the crossing velocity vector, the angle between the crossing direction and the boundary normal, and continuously track the target's movement trajectory within the restricted area in each subsequent frame. Summarize this information into an explicit prohibition trigger feature record and store it in the explicit prohibition trigger feature set.

[0073] For targets identified as being within a Level 3 defense zone, record the target number, current coordinates, current velocity vector, vertical distance from the boundary line, and the angle between the velocity direction and the boundary normal direction. Based on the current motion state, use a linear motion extrapolation model to predict the time when the target is expected to cross the boundary and the coordinates of the expected crossing point. Summarize these into a hidden trigger prohibition feature record and store it in the hidden trigger prohibition feature set.

[0074] For each feature in the explicit prohibited trigger feature set, starting from the target's current position inside the restricted area and using the current velocity vector as the initial motion condition, a uniform acceleration motion model is used to extrapolate and simulate its continuous trajectory within the restricted area. During the extrapolation process, the minimum distance from the predicted position to the nearest energized equipment is calculated in real time. When the distance between the predicted position and the energized equipment is less than the preset safe clearance at a certain moment, the complete trajectory from the moment of boundary crossing to that dangerous moment is recorded as the explicit prohibited trigger accident path.

[0075] For each feature in the set of hidden no-trigger features, starting from the target's current position, first, the approach trajectory from the current defense zone position to the boundary line is deduced, and then the subsequent trajectory after crossing the boundary and entering the restricted area is deduced. The two trajectories are connected to form a complete hidden no-trigger incident path. While predicting the path, the expected occurrence time, expected impact range, and equipment numbers involved for each incident path are recorded.

[0076] All explicit and implicit prohibited accident paths are spatiotemporally correlated and fused. Using the absolute time axis as the horizontal axis and the station's two-dimensional spatial plane as the vertical axis, all accident paths are overlaid and cross-analyzed within the same spatiotemporal coordinate system. For two or more accident paths that are less than a preset time window in time and less than a preset spatial radius in space, a spatiotemporal correlation is determined, and they are grouped into the same accident path cluster. For each accident path cluster, the top event is any target entering the prohibited constraint layer and triggering a safety accident, and the triggering conditions corresponding to each accident path in the cluster are the basic events. A prohibited accident tree is constructed using logical AND-OR gates. When a cluster contains multiple independent accident paths, the triggering of any path may lead to the top event; therefore, the basic events are related by OR gates. When a single accident path requires multiple conditions to be met simultaneously to trigger an accident, these conditions are related by AND gates.

[0077] For each node in the fault tree, multi-node suppression parameters are configured, tiered according to the node's hierarchical depth and risk urgency within the fault tree. For latent trigger nodes located outside the warning buffer zone, first-level suppression parameters are configured, including warning trigger distance, vibration alert mode, and visual indicator arrow direction. For latent trigger nodes located inside the warning buffer zone near the boundary line, second-level suppression parameters are configured, including voice broadcast content template, color and frequency of audible and visual alarms, and continuous vibration intensity of the wristband. For explicit trigger nodes located at the boundary line, third-level suppression parameters are configured, including target speed values ​​for vehicle active speed limiting, deceleration values ​​for automatic braking intervention, and trigger distance for personnel safety rope winch braking. For explicit trigger nodes located deep within the restricted area, fourth-level suppression parameters are configured, including vehicle forced engine shutdown and locking commands, station-wide broadcast alarm trigger conditions, and urgency level of dispatch center pop-up alarms. All node-configured suppression parameters are summarized according to the fault tree's logical structure to form a complete suppression parameter configuration scheme, which is then encapsulated as a third-level operation and maintenance strategy output.

[0078] For each target in the converter station monitoring visualization model, its position in the converter station's zoning constraint map is queried in real time. The target's current coordinates are read, and spatial inclusion is determined by comparing these coordinates with the three-layer spatial boundaries in the zoning constraint map. If the target coordinates are within the work constraint layer polygon defined by a work order, and the target exists in the list of permitted personnel or vehicles for that work order, then the target is marked as a work layer target, and the first maintenance strategy is applied to it. If the target coordinates are within a work-independent secondary area, and the target holds valid on-site access qualifications but does not hold a work order covering its current area, then the target is marked as an irrelevant layer target, and the second maintenance strategy is applied to it. If the target coordinates are within a work-independent primary area, or within the warning buffer zone outside the prohibited constraint layer and its movement trend points into the prohibited area, then the target is marked as a prohibited layer associated target, and the third maintenance strategy is applied to it.

[0079] For the execution of the first maintenance strategy, parameters such as speed limit values, heading correction amounts, and pause durations are encapsulated according to target type into speed limit commands and steering guidance commands sent to the vehicle's onboard controller, and work reminders and collaborative early warning information sent to personnel's smart wristbands or mobile terminals. For the execution of the second maintenance strategy, parameters such as the trigger distance threshold for tiered early warnings, voice broadcast content templates, and vibration intensity levels are encapsulated into deviation warning commands sent to wristbands or onboard terminals. For the execution of the third maintenance strategy, parameters such as the trigger distance for tiered early warnings, speed limit target values, braking intervention timing, and engine shutdown lockout conditions are encapsulated into hard interception commands sent to the vehicle controller and emergency alarm commands sent to the station's broadcast system.

[0080] The encapsulated execution instructions are distributed to the corresponding execution terminals via the industrial wireless network. A feedback confirmation mechanism for instruction execution is established. After receiving the instruction and completing the execution action, each type of execution terminal sends back an execution confirmation signal, including a successful instruction reception indicator, action execution status, and real-time status data after execution. The real-time location of all targets, constraint layer affiliation, currently applied strategy type, and instruction execution status are synchronously displayed on the converter station monitoring visualization model on the large screen of the dispatch center.

[0081] As attached Figure 2The figure shows the dynamic evolution and risk reduction curves of converter station operations. The horizontal axis represents the operation time (in hours), and the vertical axis represents the risk level and reduction rate (in %). Curve A represents the risk level of collaborative operations, curve B represents the risk level of operation deviation, curve C represents the risk level of prohibited triggering, and curve D represents the overall risk reduction rate. The figure shows the dynamic changes of the three types of risks during the operation period from 09:00 to 17:00, as well as the overall risk reduction effect after the system implemented dynamic operation and maintenance strategies. 14:00-15:00 is the peak risk period, with the risk level of collaborative operations reaching 85%, the risk level of operation deviation reaching 45%, and the risk level of prohibited triggering reaching 25%. After implementing the three-level dynamic operation and maintenance strategy, the overall risk reduction rate gradually increased from 15% to 70%.

[0082] In summary, the multi-platform integrated converter station operator and vehicle collaborative positioning method provided in this application has the following technical effects: by constructing a three-layer partition constraint map jointly driven by access data and work tickets, and deeply coupling it with a visualization model that integrates video-positioning dynamic calculation, multi-level intelligent decision-making is configured for collaborative operation accident suppression, operation deviation suppression, and prohibited trigger suppression for the work constraint layer, irrelevant constraint layer, and prohibited constraint layer, respectively. This enables hierarchical and precise collaborative management and control of personnel and vehicles and proactive intervention in accidents, thereby improving the operation and maintenance efficiency of converter stations.

[0083] Example 2: Based on the same inventive concept as the multi-platform integrated converter station personnel and vehicle collaborative positioning method in Example 1, this application also provides a multi-platform integrated converter station personnel and vehicle collaborative positioning system. Please refer to the appendix. Figure 3The multi-platform integrated converter station personnel and vehicle collaborative positioning system includes: a data acquisition module 11, used to acquire access personnel and vehicle data, work permit data, positioning monitoring data, and video surveillance data of the converter station through multiple platforms; a constraint graph construction module 12, used to construct a converter station partition constraint graph based on the access personnel and vehicle data and the work permit data, the converter station partition constraint graph including a personnel and vehicle work constraint layer, a personnel and vehicle irrelevant constraint layer, and a personnel and vehicle prohibition constraint layer; and a dynamic calculation and fusion module 13, used to perform dynamic calculation and fusion on the positioning monitoring data and the video surveillance data to obtain a converter station monitoring visualization model; and collaborative operation... The operation accident suppression configuration module 14 is used to configure collaborative operation accident suppression for the converter station monitoring visualization model according to the personnel and vehicle work constraint layer, and obtain a first operation and maintenance strategy; the operation deviation accident suppression configuration module 15 is used to configure operation deviation accident suppression for the converter station monitoring visualization model according to the personnel and vehicle irrelevant constraint layer, and obtain a second operation and maintenance strategy; the prohibition triggering accident suppression configuration module 16 is used to configure prohibition triggering accident suppression for the converter station monitoring visualization model according to the personnel and vehicle prohibition constraint layer, and obtain a third operation and maintenance strategy, and perform dynamic operation and maintenance of the converter station in combination with the first operation and maintenance strategy and the second operation and maintenance strategy.

[0084] Furthermore, the data acquisition module 11 in the multi-platform integrated converter station operator and vehicle collaborative positioning system is also used to: generate access control instructions based on the access personnel and vehicle data, and execute the dynamic access control management of the converter station based on the access control instructions.

[0085] Furthermore, the constraint map construction module 12 in the multi-platform integrated converter station operator and vehicle collaborative positioning system is also used for: dividing the converter station according to the work order data, determining work-related areas and work-independent areas; performing risk classification on the work-independent areas to obtain work-independent primary areas and work-independent secondary areas; performing time-division mapping and binding on the work-related areas according to the access personnel and vehicle data and the work order data to obtain the personnel and vehicle work constraint layer; performing time-division mapping and binding on the work-independent secondary areas according to the access personnel and vehicle data and the work order data to obtain the personnel and vehicle unrelated constraint layer; performing time-division mapping and binding on the work-independent primary areas according to the access personnel and vehicle data and the work order data to obtain the personnel and vehicle prohibition constraint layer; and combining the personnel and vehicle work constraint layer and the personnel and vehicle unrelated constraint layer to generate the converter station partition constraint map.

[0086] Furthermore, the dynamic calculation and fusion module 13 in the multi-platform integrated converter station operator and vehicle collaborative positioning system is also used for: performing occlusion and blur detection on the video monitoring data to obtain a video occlusion and blur feature set; performing interpolation compensation on the video monitoring data based on the video occlusion and blur feature set to obtain optimized monitoring data; performing dynamic calculation on the positioning monitoring data based on the optimized monitoring data to obtain positioning calculation results; and performing fusion modeling based on the optimized monitoring data and the positioning calculation results to generate the converter station monitoring visualization model.

[0087] Furthermore, the dynamic calculation and fusion module 13 in the multi-platform integrated converter station operator and vehicle collaborative positioning system is also used for: reading the terminal positioning K-th parameter of the K-th target based on the positioning monitoring data, and reading the video positioning K-th parameter of the K-th target based on the optimized monitoring data, where K is a positive integer; calculating the K-th positioning deviation based on the terminal positioning K-th parameter and the video positioning K-th parameter; determining whether the K-th positioning deviation is greater than or equal to the positioning deviation threshold; if the K-th positioning deviation is greater than or equal to the positioning deviation threshold, performing trajectory tracking on the optimized monitoring data based on the video positioning K-th parameter to obtain the tracking result of the K-th target; performing correlation correction on the terminal positioning K-th parameter based on the tracking result of the K-th target to obtain the K-th positioning calculation parameter, and adding the K-th positioning calculation parameter to the positioning calculation result.

[0088] Furthermore, the dynamic calculation and fusion module 13 in the multi-platform integrated converter station operator and vehicle collaborative positioning system is also used to: if the Kth positioning deviation is less than the positioning deviation threshold, fuse the terminal positioning Kth parameter and the video positioning Kth parameter, and output the Kth positioning calculation parameter.

[0089] Furthermore, the collaborative operation accident suppression configuration module 14 in the multi-platform integrated converter station operator and vehicle collaborative positioning system is also used for: performing spatiotemporal mapping differential comparison on the converter station monitoring visualization model according to the personnel and vehicle work constraint layer to generate a collaborative operation differential itemset; performing accident deduction based on the collaborative operation differential itemset to construct a collaborative operation accident evolution chain with collaborative operation accident risk degree; configuring a collaborative operation accident suppression decision set based on the collaborative operation accident evolution chain; performing accident suppression simulation on the converter station monitoring visualization model based on the collaborative operation accident suppression decision set to obtain simulation data for each suppression decision; performing accident risk reduction calculation on the simulation data for each suppression decision based on the collaborative operation accident risk degree to obtain each accident risk reduction degree; filtering the collaborative operation accident suppression decision set based on each accident risk reduction degree, and outputting the first operation and maintenance strategy.

[0090] Furthermore, the operation deviation accident suppression configuration module 15 in the multi-platform integrated converter station operator and vehicle collaborative positioning system is also used for: mapping the personnel and vehicle-independent constraint layer to the converter station monitoring visualization model, identifying explicit deviation triggering feature sets and implicit deviation triggering feature sets; evaluating the risk level of the explicit deviation triggering feature sets and the implicit deviation triggering feature sets respectively, and obtaining a deviation triggering risk level set; performing attention configuration fusion on the explicit deviation triggering feature sets and the implicit deviation triggering feature sets according to the deviation triggering risk level set, and obtaining an explicit-implicit deviation triggering coupling graph; configuring explicit deviation regression constraints based on the explicit deviation triggering feature sets, and configuring implicit deviation regression constraints based on the implicit deviation triggering feature sets; configuring accident suppression parameters on the explicit-implicit deviation triggering coupling graph based on the explicit deviation regression constraints and the implicit deviation regression constraints, and generating the second operation and maintenance strategy.

[0091] Furthermore, the accident suppression configuration module 16 in the multi-platform integrated converter station operator and vehicle collaborative positioning system is also used to: map the personnel and vehicle prohibition constraint layer to the converter station monitoring visualization model, identify explicit and implicit accident prohibition feature sets; predict explicit accident prohibition paths based on the explicit accident prohibition feature sets, and predict implicit accident prohibition paths based on the implicit accident prohibition feature sets; perform spatiotemporal correlation fusion on the explicit and implicit accident prohibition paths to construct an accident prohibition tree; configure multi-node suppression parameters on the accident prohibition tree to generate the third operation and maintenance strategy.

[0092] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The multi-platform integrated converter station operator and vehicle collaborative positioning method and specific examples in the aforementioned embodiment one are also applicable to the multi-platform integrated converter station operator and vehicle collaborative positioning system of this embodiment. Through the foregoing detailed description of the multi-platform integrated converter station operator and vehicle collaborative positioning method, those skilled in the art can clearly understand the multi-platform integrated converter station operator and vehicle collaborative positioning system of this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0093] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0094] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A multi-platform integrated method for collaborative positioning of personnel and vehicles in converter stations, characterized in that, include: Data on personnel and vehicles entering the converter station, work permits, location monitoring data, and video surveillance data are obtained through multiple platforms. Based on the access personnel and vehicle data and the work ticket data, a converter station zoning constraint map is constructed. The converter station zoning constraint map includes a personnel and vehicle work constraint layer, a personnel and vehicle irrelevant constraint layer, and a personnel and vehicle prohibition constraint layer. Dynamic calculation and fusion are performed on the location monitoring data and the video surveillance data to obtain a converter station monitoring visualization model; Based on the personnel and vehicle work constraint layer, the converter station monitoring visualization model is configured with collaborative operation accident suppression to obtain the first operation and maintenance strategy. Based on the personnel and vehicle-independent constraint layer, the converter station monitoring visualization model is configured to suppress operational deviation accidents, and a second operation and maintenance strategy is obtained. Based on the personnel and vehicle prohibition constraint layer, the monitoring and visualization model of the converter station is configured to prevent accident triggering, and a third operation and maintenance strategy is obtained. The dynamic operation and maintenance of the converter station is then executed in combination with the first operation and maintenance strategy and the second operation and maintenance strategy.

2. The multi-platform integrated converter station operator and vehicle collaborative positioning method as described in claim 1, characterized in that, Based on the access personnel and vehicle data and the work permit data, a converter station zoning constraint graph is constructed. The converter station zoning constraint graph includes a personnel and vehicle work constraint layer, a personnel and vehicle irrelevant constraint layer, and a personnel and vehicle prohibition constraint layer, including: The converter station is divided according to the work order data to determine the work-related area and the work-unrelated area; The risk classification of the work-irrelevant areas is performed to obtain the primary work-irrelevant areas and the secondary work-irrelevant areas. Based on the access personnel and vehicle data and the work ticket data, the work-related region is time-divisionally mapped and bound to obtain the personnel and vehicle work constraint layer; Based on the access personnel and vehicle data and the work ticket data, the work-independent secondary area is time-division mapped and bound to obtain the personnel and vehicle-independent constraint layer; Based on the access personnel and vehicle data and the work ticket data, the work-irrelevant primary area is time-divisionally mapped and bound to obtain the personnel and vehicle prohibition constraint layer. The converter station partition constraint map is generated by combining the personnel and vehicle work constraint layer and the personnel and vehicle irrelevant constraint layer.

3. The multi-platform integrated converter station personnel and vehicle collaborative positioning method as described in claim 1, characterized in that, Dynamic calculation and fusion are performed on the location monitoring data and the video surveillance data to obtain a converter station monitoring visualization model, including: Perform occlusion and blur detection on the video surveillance data to obtain a video occlusion and blur feature set; The video surveillance data is interpolated and compensated based on the video occlusion blur feature set to obtain optimized surveillance data. The location monitoring data is dynamically calculated based on the optimized monitoring data to obtain the location calculation result; The converter station monitoring visualization model is generated by fusing the optimized monitoring data and the positioning calculation results into a fusion model.

4. The multi-platform integrated converter station operator and vehicle collaborative positioning method as described in claim 3, characterized in that, The location monitoring data is dynamically calculated based on the optimized monitoring data to obtain the location calculation result, including: The terminal positioning parameter K of the Kth target is read based on the positioning monitoring data, and the video positioning parameter K of the Kth target is read based on the optimized monitoring data, where K is a positive integer; Calculate the Kth positioning deviation based on the terminal positioning Kth parameter and the video positioning Kth parameter; Determine whether the Kth positioning deviation is greater than or equal to the positioning deviation threshold; If the Kth positioning deviation is greater than or equal to the positioning deviation threshold, the optimized monitoring data is tracked according to the Kth video positioning parameter to obtain the Kth target tracking result. Based on the tracking result of the Kth target, the Kth parameter of the terminal positioning is correlated and corrected to obtain the Kth positioning solution parameter, and the Kth positioning solution parameter is added to the positioning solution result.

5. The multi-platform integrated converter station personnel and vehicle collaborative positioning method as described in claim 1, characterized in that, Based on the personnel and vehicle work constraint layer, the converter station monitoring visualization model is configured with collaborative operation accident suppression to obtain a first operation and maintenance strategy, including: Based on the personnel and vehicle work constraint layer, the spatiotemporal mapping differential comparison of the converter station monitoring visualization model is performed to generate a collaborative operation differential itemset. Accident simulation is performed based on the collaborative operation differential itemset to construct a collaborative operation accident evolution chain with collaborative operation accident risk level; Configure a collaborative operation accident suppression decision set based on the aforementioned collaborative operation accident evolution chain; Based on the collaborative operation accident suppression decision set, the converter station monitoring visualization model is used to perform accident suppression simulation to obtain simulation data for each suppression decision. Based on the collaborative operation accident risk level, the accident risk reduction calculation is performed on each suppression decision simulation data to obtain the accident risk reduction degree. Based on the risk reduction of each accident, the collaborative operation accident suppression decision set is selected, and the first operation and maintenance strategy is output.

6. The multi-platform integrated converter station personnel and vehicle collaborative positioning method as described in claim 1, characterized in that, Based on the personnel and vehicle-independent constraint layer, the converter station monitoring visualization model is configured to suppress operational deviation accidents, and a second operation and maintenance strategy is obtained, including: The personnel and vehicle-independent constraint layer is mapped to the converter station monitoring visualization model to identify explicit deviation trigger feature sets and implicit deviation trigger feature sets; The risk levels of the explicit deviation triggering feature set and the implicit deviation triggering feature set are evaluated respectively to obtain the deviation triggering risk level set; Based on the deviation trigger risk level set, attention configuration fusion is performed on the explicit deviation trigger feature set and the implicit deviation trigger feature set to obtain an explicit-implicit deviation trigger coupling graph; Explicit deviation regression constraints are configured based on the explicit deviation triggering feature set, and implicit deviation regression constraints are configured based on the implicit deviation triggering feature set. Based on the explicit deviation regression constraints and the implicit deviation regression constraints, the accident suppression parameters are configured on the explicit and implicit deviation triggering coupling graph to generate the second operation and maintenance strategy.

7. The multi-platform integrated converter station personnel and vehicle collaborative positioning method as described in claim 1, characterized in that, Based on the personnel and vehicle prohibition constraint layer, the converter station monitoring visualization model is configured to prevent accident triggering and suppress them, thereby obtaining a third operation and maintenance strategy, including: The personnel and vehicle prohibition constraint layer is mapped to the converter station monitoring visualization model to identify explicit prohibition trigger feature sets and implicit prohibition trigger feature sets; Based on the explicit prohibited trigger feature set, predict the explicit prohibited trigger accident path, and based on the implicit prohibited trigger feature set, predict the implicit prohibited trigger accident path. Spatiotemporal correlation fusion is performed on the explicit prohibited triggering accident paths and the implicit prohibited triggering accident paths to construct a prohibited triggering accident tree; Configure multi-node suppression parameters for the prohibited trigger fault tree to generate the third operation and maintenance strategy.

8. The multi-platform integrated converter station personnel and vehicle collaborative positioning method as described in claim 1, characterized in that, Data on personnel and vehicles entering the converter station, work permits, location monitoring, and video surveillance are obtained through multiple platforms, including: Based on the access personnel and vehicle data, an access control command is generated, and dynamic access control management of the converter station is executed according to the access control command.

9. The multi-platform integrated converter station personnel and vehicle collaborative positioning method as described in claim 4, characterized in that, Determining whether the Kth positioning deviation is greater than or equal to the positioning deviation threshold includes: If the Kth positioning deviation is less than the positioning deviation threshold, the Kth terminal positioning parameter and the Kth video positioning parameter are fused together to output the Kth positioning solution parameter.

10. A multi-platform integrated converter station personnel and vehicle collaborative positioning system, characterized in that, The steps for implementing the multi-platform integrated converter station personnel and vehicle collaborative positioning method according to any one of claims 1 to 9, wherein the multi-platform integrated converter station personnel and vehicle collaborative positioning system comprises: The data acquisition module is used to acquire access personnel and vehicle data, work permit data, location monitoring data, and video surveillance data of the converter station through multiple platforms; The constraint graph construction module is used to construct a converter station partition constraint graph based on the access personnel and vehicle data and the work ticket data. The converter station partition constraint graph includes a personnel and vehicle work constraint layer, a personnel and vehicle irrelevant constraint layer, and a personnel and vehicle prohibition constraint layer. The dynamic calculation and fusion module is used to perform dynamic calculation and fusion on the positioning monitoring data and the video surveillance data to obtain a monitor visualization model of the converter station. The collaborative operation accident suppression configuration module is used to configure collaborative operation accident suppression on the converter station monitoring visualization model according to the personnel and vehicle work constraint layer, and obtain the first operation and maintenance strategy. The operation deviation accident suppression configuration module is used to configure the operation deviation accident suppression configuration of the converter station monitoring visualization model according to the personnel and vehicle-independent constraint layer, and obtain the second operation and maintenance strategy. The accident suppression configuration module is used to configure the accident suppression of the converter station monitoring visualization model according to the personnel and vehicle prohibition constraint layer, obtain the third operation and maintenance strategy, and perform dynamic operation and maintenance of the converter station in combination with the first operation and maintenance strategy and the second operation and maintenance strategy.

Images