Rail transit anomaly detection method and device based on digital twinborn technology

Abstract

The invention discloses a rail transit anomaly detection method and device based on a digital twin technology, and relates to the field of artificial intelligence, the field of big data or other related fields, and the detection method comprises the steps: collecting environment perception data and video monitoring data of at least one target road section, and obtaining a real-time data stream in a first target time period; analyzing the real-time data flow based on a preset abnormal behavior rule to obtain an abnormal analysis result; a digital twinborn model is called, the digital twinborn model is detected according to the abnormal analysis result and the real-time data flow, a detection result is obtained, and the digital twinborn model is used for synchronously reflecting the actual physical environment, the basic equipment layout and the equipment operation state in the rail transit scene; and generating an anomaly detection report based on the detection result and triggering an anomaly warning. The technical problem that the emergency response speed is low due to the fact that a traditional video monitoring system lacks intelligent means in on-orbit traffic elevated section safety supervision in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence technology and big data technology, and specifically to a rail transit anomaly detection method and device based on digital twin technology. Background Art

[0002] With the rapid advancement of urbanization, the safety monitoring and operational management of elevated rail transit sections are facing challenges. While traditional monitoring methods have maintained the daily operation order of rail transit to a certain extent, their increasingly apparent limitations cannot be ignored.

[0003] Currently, traditional video surveillance systems provide security managers with intuitive monitoring tools by capturing and recording on-site footage in real time. However, with the continuous development of rail transit systems and increasing operational demands, traditional video surveillance systems are gradually becoming limited. For example, they often fail to quickly and accurately locate incidents in the face of emergencies, thus hindering the efficiency and effectiveness of emergency response. Furthermore, traditional surveillance systems lack the ability to intelligently identify and provide early warnings for different types of events, making it difficult to promptly detect and effectively prevent potential risks.

[0004] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0005] The embodiments of the present invention provide a rail transit anomaly detection method and device based on digital twin technology, so as to at least solve the technical problem in the related art that the traditional video surveillance system lacks intelligent means in the safety supervision of rail transit elevated sections, resulting in slow emergency response speed.

[0006] According to one aspect of an embodiment of the present invention, a rail transit anomaly detection method based on digital twin technology is provided, including: collecting environmental perception data and video surveillance data of at least one target section to obtain a real-time data stream within a first target time period; performing an anomaly analysis on the real-time data stream of the target section within the first target time period based on preset abnormal behavior rules to obtain an anomaly analysis result; calling a digital twin model, and detecting the digital twin model based on the anomaly analysis result and the real-time data stream to obtain a detection result, wherein the digital twin model is a pre-built virtual model for synchronously reflecting the actual physical environment, basic equipment layout and equipment operation status in the rail transit scene, and the detection result is used to record the location data and impact range of the abnormal event; generating an anomaly detection report for the target section within the first target time period based on the location data and impact range of the abnormal event, and triggering an anomaly warning based on the anomaly detection report.

[0007] Furthermore, the rail transit anomaly detection method based on digital twin technology also includes: based on the anomaly analysis results and real-time data stream within the first target time period, calling the digital twin model to simulate the operation of the target section within the second target time period to obtain a first simulation result, wherein the second target time period is a future time period after the first target time period; analyzing the simulated operation status recorded in the first simulation result based on the preset abnormal behavior rules to determine the probability value of an abnormal event occurring in the target section within the second target time period; when the probability value is not zero, obtaining the second simulation result of the digital twin model, wherein the second simulation result includes: the probability value, time prediction data and spatial prediction data; generating an anomaly prediction report for the target section within the second target time period based on the second simulation result, and triggering an anomaly warning based on the anomaly prediction report.

[0008] Furthermore, the rail transit anomaly detection method based on digital twin technology also includes: marking the anomaly range on the map in the digital twin model based on the position data or spatial prediction data determined by the digital twin model, and / or marking abnormal equipment points on the three-dimensional model in the digital twin model, wherein the rail transit scene is recorded in a two-dimensional form in the map, and the three-dimensional model records the rail transit scene in a three-dimensional form; and displaying the map and / or three-dimensional model after the anomaly marking is completed.

[0009] Furthermore, after collecting environmental perception data and video surveillance data of at least one target road section, it also includes: performing data preprocessing on all collected data, wherein the data preprocessing includes at least the following operations: data cleaning operation and data fusion operation; wherein the data cleaning operation includes at least any one of the following: removing noise, removing duplicate data, and removing invalid data; the data fusion operation includes: performing time synchronization and spatial matching on the environmental perception data and the video surveillance data to form a consistent data set.

[0010] Furthermore, the data preprocessing also includes: data mining operations, and the data mining operations include: analyzing the consistent data set obtained after fusion to obtain a key feature set, wherein the key feature set includes abnormal behavior characteristics and structural stress change characteristics in the target road section.

[0011] Furthermore, the steps of constructing the digital twin model include: using a geographic information system to obtain geographic information of all rail transit sections within the target range, and constructing a basic geographic framework for all rail transit sections within the target range based on the geographic information; using a building information model to obtain basic equipment information for all rail transit sections within the target range, and constructing a three-dimensional equipment model of all basic equipment in all rail transit sections within the target range based on the basic equipment information; fusing the basic geographic framework and the three-dimensional equipment model to obtain the digital twin model corresponding to all rail transit sections within the target range.

[0012] Furthermore, after obtaining the digital twin model, it also includes: obtaining real-time data collected by all the basic equipment within the target range through Internet of Things technology; and updating the operating status parameters of the corresponding virtual equipment in the digital twin model based on the real-time data.

[0013] Furthermore, the step of fusing the basic geographic framework and the three-dimensional device model includes: extracting the location information of each basic device recorded in the basic geographic framework, wherein the location information includes at least one of the following: coordinate data, detection direction and detection coverage; drawing a map of all rail transit sections within the target range based on the basic geographic framework, and calibrating the virtual location information of the corresponding virtual device based on the coordinate data of each basic device; establishing a three-dimensional stereo model of all rail transit sections within the target range based on the calibrated map, and calibrating the three-dimensional device model corresponding to each virtual device in the three-dimensional stereo model based on the detection direction and the detection coverage.

[0014] According to another aspect of an embodiment of the present invention, a rail transit anomaly detection device based on digital twin technology is also provided, including: an acquisition unit, used to collect environmental perception data and video surveillance data of at least one target section to obtain a real-time data stream within a first target time period; an anomaly analysis unit, used to perform an anomaly analysis on the real-time data stream of the target section within the first target time period based on preset abnormal behavior rules to obtain an anomaly analysis result; a detection unit, used to call a digital twin model, and detect the digital twin model based on the anomaly analysis result and the real-time data stream to obtain a detection result, wherein the digital twin model is a pre-built virtual model for synchronously reflecting the actual physical environment, basic equipment layout and equipment operation status in the rail transit scene, and the detection result is used to record the location data and impact range of the abnormal event; a warning unit, used to generate an anomaly detection report of the target section within the first target time period based on the location data and impact range of the abnormal event, and trigger an anomaly warning based on the anomaly detection report.

[0015] Furthermore, the rail transit anomaly detection device based on digital twin technology also includes: a simulation module, which is used to call the digital twin model to simulate the operation of the target section in the second target time period based on the anomaly analysis results and real-time data stream in the first target time period to obtain a first simulation result, wherein the second target time period is a future time period after the first target time period; an analysis module, which is used to analyze the simulated operation status recorded in the first simulation result based on the preset abnormal behavior rules to determine the probability value of an abnormal event occurring in the target section in the second target time period; a first acquisition module, which is used to obtain the second simulation result of the digital twin model when the probability value is not zero, wherein the second simulation result includes: the probability value, time prediction data and spatial prediction data; an early warning module, which is used to generate an anomaly prediction report for the target section in the second target time period based on the second simulation result, and trigger an abnormal warning based on the anomaly prediction report.

[0016] Furthermore, the rail transit anomaly detection device based on digital twin technology also includes: a labeling module, which is used to mark the abnormal range on the map in the digital twin model based on the position data or spatial prediction data determined by the digital twin model, and / or, to mark abnormal equipment points on the three-dimensional model in the digital twin model, wherein the rail transit scene is recorded in a two-dimensional form in the map, and the three-dimensional model records the rail transit scene in a three-dimensional form; a display module, which is used to display the map and / or three-dimensional model after the abnormality labeling is completed.

[0017] Furthermore, after collecting environmental perception data and video surveillance data of at least one target road section, the rail transit anomaly detection device based on digital twin technology also includes: a preprocessing module, used to perform data preprocessing on all collected data, wherein the data preprocessing includes at least the following operations: data cleaning operation and data fusion operation; wherein the data cleaning operation includes at least any one of the following: noise removal, duplicate data removal, and invalid data removal; the data fusion operation includes: time synchronization and spatial matching of the environmental perception data and the video surveillance data to form a consistent data set.

[0018] Furthermore, the data preprocessing also includes: data mining operations, and the data mining operations include: analyzing the consistent data set obtained after fusion to obtain a key feature set, wherein the key feature set includes abnormal behavior characteristics and structural stress change characteristics in the target road section.

[0019] Furthermore, the rail transit anomaly detection device based on digital twin technology also includes: a first construction module, used to use a geographic information system to obtain the geographic information of all rail transit sections within the target range, and construct a basic geographic framework for all rail transit sections within the target range based on the geographic information; a second construction module, used to use a building information model to obtain the basic equipment information of all rail transit sections within the target range, and construct a three-dimensional equipment model of all basic equipment in all rail transit sections within the target range based on the basic equipment information; a fusion module, used to fuse the basic geographic framework and the three-dimensional equipment model to obtain the digital twin model corresponding to all rail transit sections within the target range.

[0020] Furthermore, after obtaining the digital twin model, the rail transit anomaly detection device based on digital twin technology also includes: a second acquisition module, used to obtain real-time data collected by all the basic equipment within the target range through Internet of Things technology; an update module, used to update the operating status parameters of the corresponding virtual equipment in the digital twin model based on the real-time data.

[0021] Furthermore, the fusion module includes: an extraction submodule for extracting the location information of each basic device recorded in the basic geographic framework, wherein the location information includes at least one of the following: coordinate data, detection direction, and detection coverage; a first calibration submodule for drawing a map of all rail transit sections within the target range based on the basic geographic framework, and calibrating the virtual location information of the corresponding virtual device based on the coordinate data of each basic device; a second calibration submodule for establishing a three-dimensional model of all rail transit sections within the target range based on the calibrated map, and calibrating the three-dimensional device model corresponding to each virtual device in the three-dimensional model based on the detection direction and the detection coverage.

[0022] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is also provided, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned rail transit anomaly detection methods based on digital twin technology.

[0023] According to another aspect of an embodiment of the present invention, an electronic device is also provided, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned rail transit anomaly detection methods based on digital twin technology.

[0024] According to another aspect of an embodiment of the present invention, a computer program product is also provided, comprising computer instructions, wherein when the computer instructions are executed by a processor, the steps of any one of the above-mentioned methods for detecting anomalies in rail transit based on digital twin technology are implemented.

[0025] In the present invention, a rail transit anomaly detection method based on digital twin technology is proposed. The environmental perception data and video surveillance data of at least one target section are first collected to obtain a real-time data stream within a first target time period. Then, based on preset abnormal behavior rules, an abnormality analysis is performed on the real-time data stream of the target section within the first target time period to obtain an abnormality analysis result. Then, a digital twin model is called, and the digital twin model is tested based on the abnormality analysis result and the real-time data stream to obtain a test result. The digital twin model is a pre-built virtual model for synchronously reflecting the actual physical environment, basic equipment layout and equipment operation status in the rail transit scene. The test result is used to record the location data and impact range of the abnormal event. Finally, an abnormality detection report of the target section within the first target time period is generated based on the location data and impact range of the abnormal event, and an abnormality warning is triggered based on the abnormality detection report.

[0026] The present invention integrates digital twin technology with real-time monitoring data stream analysis to construct a high-precision virtual model and integrates environmental perception and video surveillance data for intelligent anomaly detection. This significantly improves the accuracy and response efficiency of abnormal event identification in elevated rail transit sections, achieving a transition from two-dimensional planar monitoring to three-dimensional early warning, and building an intelligent, visual decision support system. Compared to existing monitoring systems where abnormal event monitoring in elevated rail transit sections is limited to single-image analysis, making it difficult to quickly locate abnormal spatial coordinates and proactively warn of the development of events, the present invention integrates digital twin technology to reflect the physical and equipment status of elevated sections in real time, facilitating in-depth mining of abnormal behavior patterns in video surveillance and environmental perception data. Combining pre-set anomaly rules and machine learning algorithms, the present invention enables the immediate identification and precise location of abnormal events. Furthermore, the present invention utilizes digital twin models and big data analysis to predict the impact range and future evolution trends of abnormal events, providing rail transit operation and maintenance personnel with a comprehensive and accurate decision support system. This significantly optimizes the emergency response process and improves processing efficiency, thereby resolving the technical problem of traditional video surveillance systems in the safety supervision of elevated rail transit sections, which lack intelligent means and lead to slow emergency response. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0028] Figure 1 is a flowchart of an optional rail transit anomaly detection method based on digital twin technology according to an embodiment of the present invention;

[0029] Figure 2 is a flow chart of an optional data collection method according to an embodiment of the present invention;

[0030] Figure 3 is a flow chart of an optional data preprocessing method according to an embodiment of the present invention;

[0031] Figure 4 is a flowchart of an optional model building method according to an embodiment of the present invention;

[0032] Figure 5 This is a flowchart of an optional rail transit abnormality warning method based on digital twin technology according to an embodiment of the present invention;

[0033] Figure 6 is a flow chart of an optional early warning monitoring method according to an embodiment of the present invention;

[0034] Figure 7 is a schematic diagram of an optional rail transit anomaly detection device based on digital twin technology according to an embodiment of the present invention;

[0035] Figure 8 This is a structural block diagram of an electronic device for executing a rail transit anomaly detection method based on digital twin technology according to an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0037] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0038] To facilitate those skilled in the art to understand the present invention, some of the terms or nouns involved in the embodiments of the present invention are explained below:

[0039] Digital Twin Technology (DTT) utilizes physical models, sensor updates, historical data, and other information to construct virtual digital models that reflect and predict the state of physical objects in real time. This technology is used to create a virtual model of an elevated rail transit section. This model can visualize the actual physical environment, infrastructure layout, and operational status of the elevated section in real time, providing a highly accurate virtual environment for detecting abnormal events.

[0040] Internet of Things Technology (IoT) is a technology that enables data exchange and remote control between the physical and virtual worlds through various sensor devices and network connections. In this application, it is used to collect real-time data from infrastructure equipment in elevated sections and update it into a digital twin model, enabling the model to reflect the actual operating status of the equipment in real time.

[0041] BIM, or Building Information Modeling, is a digital approach for managing the entire process of building design, construction, and maintenance. It's more than just 3D modeling; it's an intelligent database containing detailed information about all building components. Each building component in a BIM model is a data object with attributes, characterized by completeness, relevance, and scalability, significantly improving the efficiency and quality of construction projects.

[0042] GIS, Geographic Information System, is a system used to capture, store, analyze, manage and display geospatial data. It supports users to combine map visualization with spatial data analysis functions to understand complex information from a geospatial perspective.

[0043] The following embodiments of the present invention can be applied to various rail transit systems, applications, and devices requiring intelligent abnormal event detection and precise spatial location calculation. They can implement an intelligent early warning system for elevated rail transit sections based on digital twin technology and multi-source data fusion. This invention uses a digital twin model for real-time monitoring of abnormal events, then combines environmental perception data with video surveillance data for in-depth analysis. This allows for better localization of abnormal events, prediction of their impact, and triggering of immediate warnings.

[0044] In specific implementation, the present invention first ensures the accuracy and real-time nature of the data by collecting environmental perception data and video surveillance data in real time; then uses preset abnormal behavior rules to intelligently analyze the real-time data stream and automatically identify abnormal behaviors and events; then calls the digital twin model, combines the analysis results with real-time data for in-depth detection, accurately calculates the spatial location of abnormal events and assesses their impact range; finally, generates an anomaly detection report and triggers a warning mechanism to ensure that rail transit operation and maintenance personnel can receive alerts in a timely manner and take swift countermeasures.

[0045] By integrating digital twin technology with real-time data stream analysis, this invention effectively addresses the limitations of traditional monitoring systems in event location and early warning, avoiding the errors of manual monitoring and the inaccuracies of computational models. It also overcomes the challenges of video analysis technology in complex environments, significantly improving the efficiency and reliability of safety supervision of elevated rail transit sections. This intelligent early warning system is not only applicable to rail transit safety monitoring, but also provides powerful technical support for urban rail transit operations management, emergency response, and intelligent decision-making, promoting the modernization of the transportation industry.

[0046] The present invention will be described in detail below with reference to various embodiments.

[0047] Example 1

[0048] According to an embodiment of the present invention, an embodiment of a rail transit anomaly detection method based on digital twin technology is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0049] Safety monitoring on elevated rail transit sections primarily relies on video surveillance and environmental sensing equipment. While these devices can detect anomalies at a macro level (such as pedestrian intrusions, equipment failures, or precursors to natural disasters), they struggle to pinpoint the specific coordinates of the event and predict its further development. This presents significant technical obstacles in emergency situations, especially those requiring immediate response. For example, when elevated sections encounter sudden extreme weather events such as strong winds or heavy rain, existing monitoring systems can only passively record the event, failing to proactively warn and provide accurate location information. This undoubtedly increases the difficulty and uncertainty of responding to emergencies.

[0050] The implementation subject of the rail transit anomaly detection method based on digital twin technology proposed in the embodiment of the present invention can be the rail transit elevated section intelligent early warning platform, or integrated into various rail transit safety-related equipment and applications such as existing rail transit monitoring systems, emergency command centers, and intelligent operation and maintenance systems. It combines digital twin technology with big data analysis, machine learning and other algorithms, as well as data cleaning, fusion, mining and other data processing technologies, and is used in rail transit elevated section safety monitoring scenarios, especially to solve the problems of inaccurate positioning of abnormal events and slow response speed, and to avoid safety hazards caused by inaccurate positioning of abnormal events or delayed processing. Through the embodiment of the present invention, the rail transit elevated section intelligent early warning platform is supported to enable rail transit operation and maintenance personnel to quickly formulate emergency response plans based on accurate spatial location information and comprehensive event analysis reports, effectively reduce accident risks, improve the stability of the rail transit system and the sense of security of passengers, and promote the rail transit industry towards a safer, smarter and more efficient future.

[0051] The embodiments of the present invention are described in detail below in conjunction with specific implementation steps.

[0052] Figure 1 is a flow chart of an optional rail transit anomaly detection method based on digital twin technology according to an embodiment of the present invention, such as Figure 1 As shown, the method includes the following steps:

[0053] Step S101: Collect environmental perception data and video surveillance data of at least one target road section to obtain a real-time data stream within a first target time period.

[0054] Specifically, the target section refers to the elevated section in the rail transit system that requires key monitoring. It is selected based on the geographical location, frequent traffic volume or past safety record required by this inspection, ensuring that the early warning system can cover the most critical operating areas and promptly detect and respond to potential threats.

[0055] The first target time period is a specific time window for data collection, which can be flexibly set based on the specific inspection requirements. For example, it can be 24-hour continuous monitoring or focused collection during peak hours. The purpose of setting the first target time period is to ensure that the system can capture environmental changes and video surveillance information during critical operating moments on the target road section, thereby improving the accuracy and timeliness of abnormal event detection.

[0056] Environmental perception data contains a series of key parameters related to the target section. It refers to real-time environmental information collected by the sensor network deployed on the elevated section of the rail transit. It reflects the real-time environmental conditions and physical status of the target section, including but not limited to temperature, humidity, wind speed, vibration, noise, etc. It is crucial for identifying abnormal events and assessing potential impacts. For example, an abnormal temperature rise may indicate overheating of the contact network equipment, while a sudden increase in wind speed may warn of a threat to the safety of the elevated bridge body.

[0057] Video surveillance data refers to real-time video streams captured by high-definition cameras and other video surveillance equipment, including but not limited to train operation information, passenger behavior information, and equipment status information. These cameras, located at key locations along the target route, such as bridges, stations, and tunnel entrances, continuously monitor the operational status of elevated rail transit sections. They not only provide real-time visual monitoring but also serve as a key source for identifying abnormal behavior, such as pedestrian intrusions and dropped objects, which can be captured and analyzed promptly through changes in video features.

[0058] Real-time data flow refers to the continuous flow of information from environmental sensing devices and video surveillance systems to the data processing center during the first target time period. It contains the latest snapshots of the environmental status and video footage of the target road section. It is crucial for the real-time calibration of the digital twin model and the analysis of immediate abnormal events. The timeliness and integrity of the real-time data flow are key factors in ensuring the performance of the early warning system. It ensures that abnormal conditions in elevated rail transit sections can be quickly identified and located, buying valuable time for emergency response.

[0059] In an embodiment of the present invention, video surveillance data is combined with environmental perception data and fused and analyzed through intelligent algorithms. Abnormal events such as train derailment, equipment failure, contact network fireworks, pipeline heating, and abnormal passenger behavior can be identified. The digital twin model is then used for detection to accurately locate the specific location of the abnormal event (such as tunnel entrance, platform, interior of the carriage, etc.) and the scope of impact (such as the number of affected trains, passenger safety area, etc.). This solves the problems of inaccurate abnormal event positioning and difficult impact range assessment in traditional rail transit monitoring systems, thereby improving the safety and operational efficiency of the rail transit system.

[0060] An optional embodiment, Figure 2is a flow chart of an optional data collection method according to an embodiment of the present invention. Figure 2 As shown, the method process includes: start → hardware connection and configuration → data acquisition program startup → concurrent acquisition of sensor and camera data → data verification and preprocessing → end. The following describes the process nodes.

[0061] 1. Hardware connection and configuration: First, ensure that the hardware deployment of the tower sensing network (including vibration, temperature, wind speed and other sensors) and the panoramic video monitoring system is completed, and establish a connection with the data acquisition server through a wired or wireless network. Configure the corresponding API interface or direct connection parameters so that the data acquisition program can correctly access and obtain data.

[0062] 2. Data Collection Program: Design and implement an efficient data collection program that can concurrently acquire data from multiple sensors and cameras in real time. Use multi-threading or asynchronous I / O techniques to improve data collection efficiency and real-time performance. The program must also include data caching and retry mechanisms to mitigate data loss caused by network fluctuations or device failures.

[0063] 3. Data verification and preprocessing: During the data collection process, the raw data is preliminarily verified and preprocessed. For example, the data integrity, format correctness and timestamp consistency are checked, and data that does not meet the requirements is marked or discarded to ensure the data quality of subsequent processing.

[0064] Optionally, after collecting environmental perception data and video surveillance data of at least one target road section, it also includes: performing data preprocessing on all collected data, wherein the data preprocessing includes at least the following operations: data cleaning operation and data fusion operation; wherein the data cleaning operation includes at least any one of the following: removing noise, removing duplicate data, and removing invalid data; the data fusion operation includes: temporally synchronizing and spatially matching the environmental perception data with the video surveillance data to form a consistent data set.

[0065] It should be noted that data preprocessing is performed after data collection and before anomaly analysis. Its purpose is to ensure the accuracy and efficiency of subsequent analysis and processing. It specifically includes two major steps: data cleaning operations and data fusion operations.

[0066] Among them, data cleaning operations are designed to eliminate noise, duplication, and invalid parts in the data, thereby improving data quality, enhancing the accuracy of subsequent data processing and abnormal event identification, reducing false alarm rates, and reducing unnecessary data processing burdens, further improving the response speed and stability of the entire system. In specific operations, noise removal refers to identifying and filtering irrelevant or meaningless data points, such as occasional faulty readings of sensors or false triggers in video surveillance; deduplication refers to checking for duplicate records in the data set, especially when multiple sensors cover the same area. This ensures that each indicator at each time point is recorded only once; invalid data removal refers to identifying and excluding invalid data caused by sensor offline, video surveillance failure, or other technical reasons, ensuring that all input data is valid and usable.

[0067] Additionally, data fusion integrates different types of environmental perception data and video surveillance data to ensure temporal and spatial consistency, forming a unified dataset. This creates consistency and correlation between the data, allowing for mutual verification between environmental perception data and video surveillance data, further improving the accuracy and robustness of abnormal event detection. Furthermore, the fused dataset is easier to analyze comprehensively, helping to discover potential event patterns and associations, and providing a solid foundation for intelligent analysis in early warning systems. Specifically, time synchronization involves calibrating the data's timestamps to ensure alignment of environmental perception data and video surveillance data on the timeline, even when the data originates from different sensors and observation points. Spatial matching involves mapping environmental perception data and video surveillance data to specific locations in the digital twin model in three-dimensional space, ensuring that the data's spatial coordinates correspond to the virtual environment in the model, and achieving seamless integration between the data and the model.

[0068] Optionally, the data preprocessing further includes: data mining operations, which include: analyzing the consistent data set obtained after fusion to obtain a key feature set, wherein the key feature set includes abnormal behavior features and structural stress change features in the target road section.

[0069] It's important to note that the purpose of data mining is to extract meaningful key features from the fused data set, particularly patterns associated with anomalous behavior and indicators of structural stress changes. Through data analysis and machine learning techniques, the system can identify potential safety hazards within massive amounts of environmental perception and video surveillance data, providing more accurate data for subsequent incident monitoring and early warning.

[0070] Among them, data mining operations involve two major steps: key feature extraction and feature set construction.

[0071] Key feature extraction involves using statistical analysis, pattern recognition, and machine learning algorithms to extract features from a dataset that are closely related to abnormal behavior and structural stress changes. For example, abnormal behavior features can include sudden temperature increases, unusual vibration patterns, or unexpected motion patterns captured in video surveillance. Structural stress change features can be reflected in subtle deformations of bridge pillars or changes in catenary tension.

[0072] Feature set construction involves organizing the extracted key features into a feature set, which serves as the basis for subsequent early warning monitoring and event identification. Feature sets encompass not only numerical features but also time series features and spatial distribution features, reflecting the evolution of abnormal events across time and space.

[0073] An optional embodiment, Figure 3 is a flow chart of an optional data preprocessing method according to an embodiment of the present invention. Figure 3 As shown in Figure 1, the method flow includes: start → run the data cleaning program → remove noise, duplicates, and invalid information → run the data fusion program → time synchronization and spatial matching → run the data mining program → extract key features → store or pass to the next module → end. The following is a detailed description of the process nodes.

[0074] Data cleaning involves designing and implementing a data cleaning program that automatically identifies and removes noise, duplication, and invalid information from the data. Data is cleaned according to pre-set cleaning rules using techniques such as rule-based filtering and cluster analysis. Abnormal data and processing results are recorded during the cleaning process for subsequent analysis and optimization.

[0075] Data fusion involves designing and implementing a data fusion program that synchronizes time and spatially aligns data from different sensors and cameras. This program uses timestamps and spatial coordinates to integrate related data into a consistent dataset. Furthermore, conflicts and inconsistencies during the data fusion process are addressed to ensure the accuracy and integrity of the fused data.

[0076] Data mining involves designing and implementing data mining programs that use machine learning algorithms (such as support vector machines and neural networks) to mine and analyze fused data, extracting key features such as abnormal behavior patterns and structural stress changes, providing strong support for subsequent early warning and location tracking. At the same time, mining algorithms are continuously optimized and updated to adapt to the ever-changing data environment and abnormal patterns.

[0077] Step S102 : performing an abnormality analysis on the real-time data stream of the target road section within the first target time period based on preset abnormal behavior rules to obtain an abnormality analysis result.

[0078] It should be noted that pre-defined abnormal behavior rules refer to a set of rules for identifying and determining abnormal events, based on historical events, expert knowledge, and data analysis. In embodiments of the present invention, these pre-defined abnormal behavior rules are combined with machine learning algorithms to analyze real-time data streams and quickly and accurately identify abnormal events.

[0079] As an example, in the embodiment of the present invention, the abnormal behavior rules that can be set include but are not limited to any of the following:

[0080] 1. Catenary fire and smoke detection, that is, monitoring whether there is abnormal smoke or fire in the video screen, and whether the temperature in the environmental perception data rises abnormally, and combining the two to determine whether there is a fire risk such as overheating or short circuit of the catenary equipment; 2. Pipeline heating warning, that is, setting the normal temperature range of the catenary equipment. Once the sensor reading exceeds this range, it is regarded as a possible abnormal precursor; 3. Wind speed mutation monitoring, that is, using wind speed sensors to monitor crosswind intensity, define a normal wind speed range, and when the value reported by the wind speed sensor significantly exceeds this range, it is regarded as an abnormality; 4. Abnormal rainfall intensity, that is, using precipitation sensors to detect whether the precipitation exceeds the maximum precipitation for safe operation of the elevated section. Rainfall events that exceed the safety threshold can be regarded as abnormal precursors; 5. Abnormal video behavior, that is, by identifying human activities or natural phenomena in the picture, such as unidentified objects entering the limit, abnormal train stops, etc., as preliminary warning signals.

[0081] It's also important to note that anomaly analysis results are warnings derived from a preliminary analysis of real-time data streams based on pre-set abnormal behavior rules. This phase focuses on the initial location and classification of anomalies, rather than the precise calculation of their spatial locations. These results primarily include a list of suspected anomalies, an identification of their type, a time stamp for each event, and a preliminary confidence score.

[0082] Among them, the suspected abnormal event list contains all suspected abnormal events that have been preliminarily identified, such as abnormal temperature, excessive wind speed, excessive rainfall, etc.; the abnormal event type identifier is used to classify each suspected abnormal event and identify which type of abnormal precursor it belongs to under the preset rules; the event time tag is used to record the specific time when each suspected abnormal event occurred, which is convenient for time series analysis and subsequent event tracing; the preliminary confidence score is used to preliminarily score the reliability of the suspected abnormal event and judge the necessity of the alarm based on the degree of match between the data stream and the preset rules.

[0083] Step S103: call the digital twin model and detect the digital twin model based on the abnormality analysis results and real-time data stream to obtain the detection results.

[0084] Among them, the digital twin model is a pre-built virtual model used to synchronously reflect the actual physical environment, basic equipment layout and equipment operating status in the rail transit scenario. The detection results are used to record the location data and impact range of abnormal events.

[0085] Specifically, the rail transit scenario is a comprehensive virtual environment, including but not limited to the elevated rail transit sections that have been recorded and monitored at all times and states. Whether in daily operations, extreme weather conditions or special events, the digital twin model constructed in this way can reflect and simulate the situation of the entire elevated rail transit section in real time, including the operating status of each car, the health status of each bridge pillar, and changes in the surrounding environment, ensuring that any possible abnormal events can be fully monitored and predicted.

[0086] Furthermore, the digital twin model is a high-fidelity virtual replica that integrates physical modeling, sensor data, historical operational data, and predictive algorithms to fully reflect the actual physical environment, infrastructure layout, and operational status of the elevated rail transit section. This model not only provides a static three-dimensional representation but also includes a dynamic operational simulation, capable of real-time updates and synchronization with the actual environment to ensure it reflects the latest rail transit scenarios. By combining BIM, GIS, and IoT technologies, it accurately depicts the physical characteristics of each elevated rail transit section and simulates the dynamic response of equipment, such as the subtle vibrations of bridges when trains pass and the impact of weather changes on the line, providing data support and scenario simulation for intelligent analysis.

[0087] It's also important to note that the actual physical environment refers to the specific geographic environment and physical characteristics of the elevated rail section, including topography, bridge structure, support column layout, contact network layout, tunnel entrances and exits, and so on. The infrastructure layout refers to the precise location and configuration of all key equipment on the elevated rail section, including signal lights, cameras, sensors, power facilities, and communications equipment. This information forms the foundation for building a digital twin model, ensuring that the model accurately reflects every detail of the actual elevated rail section, from the static physical structure to the dynamic equipment operating status.

[0088] Furthermore, equipment operating status includes the working conditions and performance parameters of all equipment on the elevated rail transit section, such as train speed, power load, and changes in sensor readings. The digital twin model receives and updates this data in real time, simulating equipment performance under different conditions and operations, providing real-time operational context for detecting abnormal events and assessing their impact.

[0089] The location data of the abnormal event output by the digital twin model in step S103 refers to the precise coordinates of the abnormal event within the virtual model, including longitude, latitude, altitude, and relative position, obtained through detection and analysis by the digital twin model. The impact range assesses the potential physical area and operational impact of the abnormal event, such as whether it affects train traffic and potential threats to the safety of surrounding equipment. This information can serve as decision-making data for rail transit operators and safety managers.

[0090] Optionally, the steps of constructing the digital twin model include: using a geographic information system to obtain the geographic information of all rail transit sections within the target range, and constructing a basic geographic framework for all rail transit sections within the target range based on the geographic information; using a building information model to obtain the basic equipment information of all rail transit sections within the target range, and constructing a three-dimensional equipment model of all basic equipment in all rail transit sections within the target range based on the basic equipment information; fusing the basic geographic framework and the three-dimensional equipment model to obtain the digital twin models corresponding to all rail transit sections within the target range.

[0091] It should be noted that the Geographic Information System (GIS) is used to provide geographic information such as the precise geographic coordinates, topography, surrounding environment, and location information of each basic equipment in the elevated rail transit section for all rail transit sections within the target range. Specifically, geographic coordinates refer to the latitude and longitude coordinates of the rail transit section, which are used for precise positioning; topography includes the terrain features, slope, elevation, etc. along the line, providing topographic information for the construction of the three-dimensional model; the surrounding environment such as buildings, rivers, roads, etc. ensures that the model can reflect the environmental factors in the actual scene; the location information of each basic equipment in the elevated rail transit section includes the coordinate data, detection direction and detection coverage of each device. The basic geographic framework refers to the geographic information framework of the rail transit scene obtained and constructed using the Geographic Information System (GIS) technology, which includes the above-mentioned geographic coordinates, topography, surrounding environment, and location information of each basic equipment in the elevated rail transit section, providing a geographic positioning basis for the construction of the digital twin model and spatial position calculation.

[0092] Optionally, geographic information collected by a geographic information system (GIS) can also cover a wide range of environmental details to enhance the integrity and authenticity of the digital twin model, including but not limited to: topography, detailed ground height and terrain features, such as valleys, hills, rivers and lakes; road network, including details of roads, sidewalks and other transportation facilities near the elevated section of the rail transit; buildings and obstacles, including the location, height and shape information of obstacles such as buildings, trees, and utility poles along the line; climate and environmental data, including average temperature, precipitation, wind direction and speed, etc., used to simulate and predict the environmental conditions of the elevated section of the rail transit; soil type and geological structure, including soil type and geological structure information along the line, used to assess the stability of the elevated section of the rail transit; underground facilities, such as the layout of pipelines, cables and other infrastructure, used for accident analysis and the formulation of emergency response plans.

[0093] Another point worth noting is that Building Information Modeling (BIM) is used to obtain basic equipment information for all rail transit sections within the target range, including the precise location, type, and status of signal lights, cameras, sensors, electrical equipment, structural pillars, etc. When constructing a 3D equipment model, the following points need to be considered: 1. Accurate equipment positioning: ensuring that the equipment location in the model matches the actual location, with minimal error; 2. Equipment type and status: The 3D equipment model must include equipment type information and equipment operating status updated based on real-time data; 3. Equipment function and role: The 3D equipment model should not only reflect the physical characteristics of the equipment, but also its role and function in rail transit operations.

[0094] The equipment information obtained through BIM technology can support the construction of a three-dimensional equipment model that includes all key equipment. It is not only spatially aligned with the geographic framework, but also functionally interconnected and compared with reality to form a dynamic equipment network.

[0095] Optionally, the step of fusing the basic geographic framework and the three-dimensional device model includes: extracting the location information of each basic device recorded in the basic geographic framework, wherein the location information includes at least one of the following: coordinate data, detection direction, and detection coverage; drawing a map of all rail transit sections within the target range based on the basic geographic framework, and calibrating the virtual location information of the corresponding virtual device based on the coordinate data of each basic device; establishing a three-dimensional stereo model of all rail transit sections within the target range based on the calibrated map, and calibrating the three-dimensional device model corresponding to each virtual device in the three-dimensional stereo model based on the detection direction and detection coverage.

[0096] It should be noted that the three-dimensional equipment model refers to a three-dimensional virtual model of the basic equipment in the elevated section of rail transit created using building information modeling (BIM) technology. It not only contains the geometric information of each basic equipment, but also is used to reflect the real-time data of the equipment's operating status. It is an important part of building a digital twin model.

[0097] In the process of obtaining a three-dimensional device model, it is first necessary to accurately extract the location information of each basic device recorded in the basic geographic framework. The coordinate data is the latitude and longitude coordinates of each device, which is the basis for locating the device in three-dimensional space. The detection direction refers to the orientation of the camera and sensor, which is used to simulate and calibrate its field of view or sensing range to ensure that the virtual device can reflect the physical characteristics of the real device. The detection coverage range is used to define the effective detection range of the device, such as the camera's field of view angle and the sensor's sensing distance, which is used to optimize the device layout and evaluate its monitoring capabilities.

[0098] Next, the extracted location information is used to draw and calibrate a map of all rail transit sections within the target range. The corresponding virtual device positions are calibrated based on the coordinate data of each infrastructure device to ensure that the virtual device's position on the map completely corresponds to its actual position. Specifically, map drawing refers to generating a map of the rail transit sections within the target range based on the data in the basic geographic framework, including the terrain, roads, buildings, and other geographical features along the route; device position calibration refers to marking the exact location of each infrastructure device on the map, including sensors, cameras, etc., to ensure that its virtual location information is consistent with its actual location information.

[0099] Then, a 3D model is created and calibrated based on the calibrated map. A 3D model of all rail transit sections within the target range is created, and the 3D device models corresponding to each virtual device in the 3D model are further calibrated based on the detection direction and detection coverage. Specifically, 3D model construction refers to the use of BIM technology combined with a calibrated map to create a 3D model containing the elevated rail transit section and its ancillary equipment. The model should include a detailed equipment layout and structural details of the elevated section. Device model calibration refers to adjusting the position and orientation of the virtual device model in the 3D model according to the actual detection direction and coverage of the device to ensure that its performance in the model is consistent with reality. For example, the camera's field of view and the sensor's sensing range should be consistent with reality.

[0100] Optionally, after obtaining the digital twin model, it also includes: obtaining real-time data collected by all basic equipment within the target range through Internet of Things technology; and updating the operating status parameters of the corresponding virtual equipment in the digital twin model based on the real-time data.

[0101] Finally, the real-time data is further synchronized to the calibrated three-dimensional model through the Internet of Things technology. That is, the field data collected by all basic equipment (including various sensors and cameras) within the target range for continuous monitoring of environmental changes, equipment status and abnormal events are obtained and synchronized to the corresponding three-dimensional equipment model in the three-dimensional model as status parameters.

[0102] Another optional embodiment, Figure 4 is a flow chart of an optional model building method according to an embodiment of the present invention, such as Figure 4 As shown, the method flow includes: start → collect real-life images and geographic information → build a three-dimensional physical model → use GIS technology to obtain geographic information, and use GIS information for positioning and calibration → design and implement data update program → obtain real-time equipment status data → update the digital twin model → end.

[0103] It should be noted that 3D physical modeling involves using WebGL and oblique photography technology to construct a 3D physical model based on the actual conditions of the elevated section. This involves collecting real-life images of the elevated section and combining them with geographic information, processing and modeling them using specialized software, to create a precise 3D model. Simultaneously, the model is optimized and adjusted, and the 3D model is positioned and calibrated using high-precision geographic coordinates and terrain information provided by GIS technology to ensure its accuracy and consistency with the actual environment, thereby guaranteeing the accuracy of spatial positioning calculations.

[0104] IoT technology updates involve updating the operational status data of actual devices into the digital twin model in real time using IoT technology. Design and implement a data update program that automatically retrieves status data from IoT devices, matches and updates the data with the corresponding devices in the digital twin model, and records any anomalies and processing results during the update process for subsequent analysis and optimization.

[0105] Step S104: Generate an abnormality detection report for the target road section within the first target time period based on the location data and impact range of the abnormal event, and trigger an abnormality warning based on the abnormality detection report.

[0106] It should be noted that the anomaly detection report includes but is not limited to: an overview of the abnormal event (the type of abnormal event, the time and location of preliminary identification), location data (the precise coordinates of the abnormal event in the digital twin model and the relative position and geographic coordinates in the actual elevated rail section), impact range assessment (describing the physical area that the abnormal event may affect, the equipment impact, and the potential impact on rail transit operations), event severity score (combining historical event data and preset impact levels to quantify the severity of the abnormal event), and recommended measures (based on the event type and impact assessment, preliminary emergency response and maintenance guidance suggestions are proposed, such as train deceleration, line closure, equipment inspection, etc.).

[0107] Triggering an abnormality warning is an emergency notification mechanism based on the content of an abnormality detection report. It aims to promptly inform decision-makers and relevant maintenance personnel of the occurrence of an abnormal event so that swift response measures can be taken. Abnormal warnings include: 1. Instant notification: Instantly send warning information to relevant personnel via SMS, email, app push, or visual network system; 2. Multi-level warning: Set a multi-level warning mechanism based on the severity and urgency of the incident to ensure that critical incidents receive an immediate and priority response; 3. Visual display: Highlight the location of the abnormal event on the digital twin model or GIS map, intuitively presenting the impact area of ​​the event; 4. Emergency response linkage: Link with the emergency response system to automatically activate emergency plans, such as adjusting train operation plans, dispatching maintenance teams, and activating backup systems, to minimize the impact of abnormal events.

[0108] Another optional embodiment, Figure 5 This is a flow chart of an optional rail transit anomaly warning method based on digital twin technology according to an embodiment of the present invention. Figure 5 As shown, the method includes the following steps:

[0109] Step S501: Based on the anomaly analysis results and real-time data stream within the first target time period, the digital twin model is invoked to simulate the operation of the target road section within a second target time period to obtain a first simulation result, wherein the second target time period is a future time period after the first target time period.

[0110] Step S502: analyzing the simulated operation status recorded in the first simulation result based on preset abnormal behavior rules to determine a probability value of an abnormal event occurring on the target road section within a second target time period;

[0111] Step S503: When the probability value is not zero, obtain a second simulation result of the digital twin model, wherein the second simulation result includes: the probability value, the time prediction data, and the space prediction data;

[0112] Step S504: Generate an abnormality prediction report for the target road section within the second target time period based on the second simulation result, and trigger an abnormality warning based on the abnormality prediction report.

[0113] It should be noted that the digital twin model can be used to simulate the operation of the target section within the second target time period, where the second target time period is the future time period immediately following the first target time period. This operation is based on historical anomaly analysis data and current real-time data streams, and uses simulation technology to predict the future operation status of the target section. The specific steps are as follows: data fusion, fusing the anomaly analysis results within the first target time period with the real-time data stream to form a comprehensive data set containing anomaly patterns and real-time status; simulating the operation status, calling the digital twin model, and performing real-time or predictive simulation based on the comprehensive data set to obtain the first simulation result, including the future operation status, equipment status, environmental factors, etc. of the target section.

[0114] The simulated operating status recorded in the first simulation results is then further analyzed based on pre-set abnormal behavior rules to determine the probability of an abnormal event occurring on the target road section within the second target time period. This process involves: applying abnormal behavior rules, comparing behavioral patterns in historical abnormal data with the first simulation results to identify potential abnormal trends; calculating event probabilities; and calculating the probability of an abnormal event occurring on the target road section within a future time period based on the degree of matching of the abnormal behavior rules and analysis of real-time data streams.

[0115] If the calculated probability value is non-zero, the digital twin model obtains a second simulation result, which includes the probability value of the abnormal event, time prediction data, and spatial prediction data. Based on the second simulation result, an abnormality prediction report for the target road section within the second target time period is generated, and an abnormality warning is triggered. Specific operations include: generating a prediction report, compiling key information from the second simulation result into an abnormality prediction report. The report should include the probability of the abnormal event, the expected time of occurrence, the predicted location, etc.; triggering an abnormality warning, sending the abnormality prediction report to relevant safety management personnel, notifying them of the possibility of an abnormal event in advance via email, text message, or system notification, so that preventive measures can be taken.

[0116] Optionally, the rail transit anomaly detection method based on digital twin technology also includes: marking the anomaly range on the map in the digital twin model based on the location data or spatial prediction data determined by the digital twin model, and / or marking abnormal equipment points on the three-dimensional model in the digital twin model, wherein the rail transit scene is recorded in two-dimensional form in the map and the three-dimensional model records the rail transit scene in three-dimensional form; and displaying the map and / or three-dimensional model after the anomaly marking is completed.

[0117] It should be noted that, based on the spatial position data or spatial prediction data determined by the digital twin model, the embodiment of the present invention can also mark the abnormal range or abnormal equipment points on the map and three-dimensional model in the digital twin model. This process includes: position data parsing, extracting position data or spatial prediction data from the abnormal event detection results, including the latitude and longitude coordinates, altitude, and relative position description of the abnormal event; abnormal range marking, marking the abnormal impact range in two-dimensional form on the map of the digital twin model by demarcating the boundaries of specific colors or patterns; abnormal equipment point marking, marking the specific location of the abnormal equipment in three-dimensional form in the three-dimensional model through highlighting, special icons or annotations. The marking of the three-dimensional model can intuitively show the relative position of the equipment in space, especially for the elevated rail sections with complex geometric structures.

[0118] After the anomaly is marked, the anomaly event marking results of the map and three-dimensional model will be displayed. An interactive two-dimensional map can be presented on the user's terminal device. The user can view detailed information about the anomaly range by zooming in, out, dragging, etc., including specific coordinates, affected area, and associated equipment. The abnormal equipment points can also be highlighted in an intuitive manner in the three-dimensional model. The user can observe the abnormal status of the equipment from different angles and levels. At the same time, the model can also display detailed information about the abnormal event, such as event type, severity, and predicted impact.

[0119] Alternatively, Figure 6 is a flow chart of an optional early warning monitoring method according to an embodiment of the present invention, such as Figure 6 As shown, the method flow includes: start → receiving digital twin model and real-time monitoring data → running spatial position calculation program → calculating the spatial position of abnormal events → running location information display program → map annotation and three-dimensional model highlighting → running alarm information report generation program → generating and sending alarm information report → recording report sending and receiving status → end.

[0120] Embodiments of the present invention can also design and implement a spatial position calculation program that can calculate the spatial position of abnormal events based on the digital twin model and real-time monitoring data. By utilizing the three-dimensional coordinate information in the digital twin model and the position information in the real-time monitoring data, when abnormal time warning information is generated in the early warning monitoring module, it is transmitted to the visual network via communication, quickly capturing the abnormal event and making a corresponding spatial positioning judgment for the abnormal event based on the three-dimensional spatial information of the digital twin. At the same time, considering factors such as the complex terrain and building obstructions in the elevated section, the positioning results can be continuously corrected and optimized through personnel corrections, further improving positioning accuracy.

[0121] Embodiments of the present invention can also design and implement a location information display program that can intuitively display the calculated spatial location information in the form of map annotations, 3D model highlights, etc. Utilizing GIS map services and 3D rendering technology, the location information of abnormal events is marked on a map and highlighted in the form of a 3D model. Detailed location description information (such as longitude, latitude, altitude, etc.) is also provided to enable relevant personnel to quickly understand the location of the abnormal event.

[0122] Embodiments of the present invention can also design and implement an alarm information report generation program that can generate detailed alarm information reports based on the location information and early warning strategies of abnormal events. The reports include key information such as event type, location, scope of impact, and recommended measures. Reports can also be sent to relevant personnel or departments via email, PDF files, or other formats, allowing them to promptly understand and address abnormal events.

[0123] Through the above steps S101 to S104, the environmental perception data and video surveillance data of at least one target section can be first collected to obtain the real-time data stream within the first target time period, and then the real-time data stream of the target section within the first target time period can be analyzed for abnormalities based on the preset abnormal behavior rules to obtain the abnormal analysis results, and then the digital twin model is called, and the digital twin model is tested based on the abnormal analysis results and the real-time data stream to obtain the test results, wherein the digital twin model is a pre-built virtual model for synchronously reflecting the actual physical environment, basic equipment layout and equipment operation status in the rail transit scene, and the test results are used to record the location data and impact range of the abnormal event, and finally an abnormal detection report of the target section within the first target time period is generated based on the location data and impact range of the abnormal event, and an abnormal warning is triggered based on the abnormal detection report.

[0124] In an embodiment of the present invention, a high-precision virtual model is constructed by integrating digital twin technology with real-time monitoring data stream analysis. This model then combines environmental perception with video surveillance data for intelligent anomaly detection, significantly improving the accuracy and response efficiency of abnormal event identification in elevated rail transit sections. This system transitions from two-dimensional monitoring to three-dimensional early warning, achieving the technical effect of building an intelligent, visual decision support system. Compared to existing monitoring systems where abnormal event monitoring in elevated rail transit sections is limited to single-image video analysis, making it difficult to quickly locate abnormal spatial coordinates and proactively warn of the development of an event, the embodiment of the present invention integrates digital twin technology to reflect the physical and equipment status of the elevated section in real time, facilitating in-depth mining of abnormal behavior patterns in video surveillance and environmental perception data. This system, combined with pre-set anomaly rules and machine learning algorithms, enables immediate identification and precise location of abnormal events. Furthermore, the present invention utilizes digital twin models and big data analysis technology to predict the impact scope and future evolution trends of abnormal events, providing rail transit operation and maintenance personnel with a comprehensive and accurate decision support system. This significantly optimizes the emergency response process and improves processing efficiency, thereby resolving the technical problem of traditional video surveillance systems in the safety supervision of elevated rail transit sections, which lack intelligent means and result in slow emergency response.

[0125] The present invention is described below in conjunction with another optional embodiment.

[0126] Example 2

[0127] A rail transit anomaly detection device based on digital twin technology provided in this embodiment includes multiple implementation units, each implementation unit corresponding to each implementation step in the above-mentioned embodiment one.

[0128] Figure 7 is a schematic diagram of an optional rail transit anomaly detection device based on digital twin technology according to an embodiment of the present invention, such as Figure 7 As shown, the device may include: a collection unit 701, an abnormality analysis unit 702, a detection unit 703, and a warning unit 704.

[0129] The collection unit 701 is configured to collect environmental perception data and video surveillance data of at least one target road section to obtain a real-time data stream within a first target time period.

[0130] The abnormality analysis unit 702 is configured to perform abnormality analysis on the real-time data stream of the target road section within the first target time period based on preset abnormal behavior rules to obtain abnormality analysis results.

[0131] The detection unit 703 is used to call the digital twin model and detect the digital twin model based on the abnormal analysis results and real-time data stream to obtain the detection results. The digital twin model is a pre-built virtual model used to synchronously reflect the actual physical environment, basic equipment layout and equipment operation status in the rail transit scene. The detection results are used to record the location data and impact range of abnormal events.

[0132] The warning unit 704 is configured to generate an abnormality detection report for the target road section within the first target time period based on the location data and the impact range of the abnormal event, and trigger an abnormality warning based on the abnormality detection report.

[0133] The above-mentioned rail transit anomaly detection device based on digital twin technology can first collect environmental perception data and video surveillance data of at least one target section through the collection unit 70 to obtain a real-time data stream within the first target time period, and then perform an anomaly analysis on the real-time data stream of the target section within the first target time period based on preset abnormal behavior rules through the anomaly analysis unit 702 to obtain an anomaly analysis result, and then call the digital twin model through the detection unit 703, and detect the digital twin model based on the anomaly analysis result and the real-time data stream to obtain a detection result, wherein the digital twin model is a pre-built virtual model for synchronously reflecting the actual physical environment, basic equipment layout and equipment operation status in the rail transit scene, and the detection result is used to record the location data and impact range of the abnormal event, and finally generate an anomaly detection report for the target section within the first target time period based on the location data and impact range of the abnormal event through the warning unit 704, and trigger an anomaly warning based on the anomaly detection report.

[0134] In an embodiment of the present invention, a high-precision virtual model is constructed by integrating digital twin technology with real-time monitoring data stream analysis. This model then combines environmental perception with video surveillance data for intelligent anomaly detection, significantly improving the accuracy and response efficiency of abnormal event identification in elevated rail transit sections. This system transitions from two-dimensional planar monitoring to three-dimensional early warning, achieving the technical effect of building an intelligent, visual decision support system. Compared to existing monitoring systems where abnormal event monitoring in elevated rail transit sections is limited to single-image video analysis, making it difficult to quickly locate abnormal spatial coordinates and proactively warn of the development of an event, the embodiment of the present invention integrates digital twin technology to reflect the physical and equipment status of the elevated section in real time, facilitating in-depth mining of abnormal behavior patterns in video surveillance and environmental perception data. This system combines pre-set anomaly rules with machine learning algorithms to achieve instant identification and precise location of abnormal events. Furthermore, the present invention utilizes digital twin models and big data analysis technology to predict the impact scope and future evolution trends of abnormal events, providing rail transit operation and maintenance personnel with a comprehensive and accurate decision support system. This system significantly optimizes the emergency response process and improves processing efficiency, thereby resolving the technical problem of traditional video surveillance systems in the related art, which lack intelligent means for safety supervision of elevated rail transit sections and leads to slow emergency response.

[0135] Optionally, the rail transit anomaly detection device based on digital twin technology also includes: a simulation module, which is used to call the digital twin model to simulate the operation of the target section in the second target time period based on the anomaly analysis results and real-time data stream in the first target time period to obtain a first simulation result, wherein the second target time period is a future time period after the first target time period; an analysis module, which is used to analyze the simulated operation status recorded in the first simulation result based on preset abnormal behavior rules to determine the probability value of an abnormal event occurring in the target section in the second target time period; a first acquisition module, which is used to obtain the second simulation result of the digital twin model when the probability value is not zero, wherein the second simulation result includes: probability value, time prediction data and spatial prediction data; an early warning module, which is used to generate an anomaly prediction report for the target section in the second target time period based on the second simulation result, and trigger an abnormality early warning based on the anomaly prediction report.

[0136] Optionally, the rail transit anomaly detection device based on digital twin technology also includes: a labeling module, which is used to mark the anomaly range on the map in the digital twin model based on the location data or spatial prediction data determined by the digital twin model, and / or, to mark abnormal equipment points on the three-dimensional model in the digital twin model, wherein the rail transit scene is recorded in two-dimensional form in the map and the three-dimensional model records the rail transit scene in three-dimensional form; a display module, which is used to display the map and / or three-dimensional model after the anomaly labeling is completed.

[0137] Optionally, after collecting environmental perception data and video surveillance data of at least one target road section, the rail transit anomaly detection device based on digital twin technology also includes: a preprocessing module for performing data preprocessing on all collected data, wherein the data preprocessing includes at least the following operations: data cleaning operation and data fusion operation; wherein the data cleaning operation includes at least any one of the following: removing noise, removing duplicate data, and removing invalid data; the data fusion operation includes: time synchronization and spatial matching of the environmental perception data and the video surveillance data to form a consistent data set.

[0138] Optionally, the data preprocessing further includes: data mining operations, which include: analyzing the consistent data set obtained after fusion to obtain a key feature set, wherein the key feature set includes abnormal behavior features and structural stress change features in the target road section.

[0139] Optionally, the rail transit anomaly detection device based on digital twin technology also includes: a first construction module, used to use a geographic information system to obtain the geographic information of all rail transit sections within the target range, and construct a basic geographic framework for all rail transit sections within the target range based on the geographic information; a second construction module, used to use a building information model to obtain the basic equipment information of all rail transit sections within the target range, and construct a three-dimensional equipment model of all basic equipment in all rail transit sections within the target range based on the basic equipment information; a fusion module, used to fuse the basic geographic framework and the three-dimensional equipment model to obtain a digital twin model corresponding to all rail transit sections within the target range.

[0140] Optionally, after obtaining the digital twin model, the rail transit anomaly detection device based on digital twin technology also includes: a second acquisition module, used to obtain real-time data collected by all basic equipment within the target range through Internet of Things technology; an update module, used to update the operating status parameters of the corresponding virtual equipment in the digital twin model based on real-time data.

[0141] Optionally, the fusion module includes: an extraction submodule for extracting the location information of each basic device recorded in the basic geographic framework, wherein the location information includes at least one of the following: coordinate data, detection direction, and detection coverage; a first calibration submodule for drawing a map of all rail transit sections within the target range based on the basic geographic framework, and calibrating the virtual location information of the corresponding virtual device based on the coordinate data of each basic device; a second calibration submodule for establishing a three-dimensional stereo model of all rail transit sections within the target range based on the calibrated map, and calibrating the three-dimensional device model corresponding to each virtual device in the three-dimensional model based on the detection direction and detection coverage.

[0142] The above-mentioned rail transit anomaly detection device based on digital twin technology can also include a processor and a memory. The above-mentioned acquisition unit 701, anomaly analysis unit 702, detection unit 703, warning unit 704, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions.

[0143] The processor includes a kernel that retrieves the corresponding program unit from memory. One or more kernels can be configured, and the digital twin model is invoked by adjusting kernel parameters. The digital twin model is then tested based on anomaly analysis results and real-time data streams. The test results are then generated, and an anomaly detection report for the target road section within a first target time period is generated based on the location data and impact range of the anomaly event. An anomaly warning is then triggered based on the anomaly detection report.

[0144] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0145] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program initialized with the following method steps: collecting environmental perception data and video surveillance data of at least one target section to obtain a real-time data stream within a first target time period; performing an abnormality analysis on the real-time data stream of the target section within the first target time period based on preset abnormal behavior rules to obtain an abnormality analysis result; calling a digital twin model, and detecting the digital twin model based on the abnormality analysis result and the real-time data stream to obtain a detection result, wherein the digital twin model is a pre-built virtual model for synchronously reflecting the actual physical environment, basic equipment layout and equipment operation status in the rail transit scene, and the detection result is used to record the location data and impact range of the abnormal event; generating an abnormality detection report for the target section within the first target time period based on the location data and impact range of the abnormal event, and triggering an abnormality warning based on the abnormality detection report.

[0146] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is also provided, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the rail transit anomaly detection methods based on digital twin technology in the above-mentioned embodiment 1.

[0147] According to another aspect of an embodiment of the present invention, an electronic device is also provided, including one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement the rail transit anomaly detection method based on digital twin technology of any one of the above-mentioned embodiments.

[0148] Figure 8 is a structural block diagram of an electronic device for executing a rail transit anomaly detection method based on digital twin technology according to an embodiment of the present invention, such as Figure 8 As shown, the electronic device may include: one or more ( Figure 8 Only one is shown) processor 802, memory 804, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0149] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the rail transit anomaly detection method and device based on digital twin technology in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned rail transit anomaly detection method based on digital twin technology. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely located relative to the processor, and these remote memories can be connected to the terminal via a network. Examples of the above-mentioned network include but are not limited to the Internet, corporate intranet, local area network, mobile communication network and combinations thereof.

[0150] It can be understood by those skilled in the art that Figure 8 The structure shown is for illustration only, and the electronic device may also be a smart phone, a tablet computer, a PDA, a mobile internet device (MID), a PAD or other terminal device. Figure 8 It does not limit the structure of the above electronic device. For example, the electronic device may also include Figure 8 More or fewer components (such as network interfaces, display devices, etc.) shown in, or with Figure 8 Different configurations shown.

[0151] A person skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0152] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0153] In the above embodiments of the present invention, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0154] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of the units can be a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0155] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0156] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0157] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, etc. Various media that can store program codes.

[0158] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A rail transit anomaly detection method based on digital twin technology, characterized in that: include: Collecting environmental perception data and video surveillance data of at least one target road section to obtain a real-time data stream within a first target time period; performing an abnormality analysis on the real-time data stream of the target road section within the first target time period based on preset abnormal behavior rules to obtain an abnormality analysis result; Calling a digital twin model and testing the digital twin model based on the anomaly analysis result and the real-time data stream to obtain a test result, wherein the digital twin model is a pre-built virtual model for synchronously reflecting the actual physical environment, basic equipment layout, and equipment operating status in the rail transit scene, and the test result is used to record the location data and impact range of the abnormal event; An abnormality detection report for the target road section within the first target time period is generated based on the location data and impact range of the abnormal event, and an abnormality warning is triggered based on the abnormality detection report.

2. The rail transit anomaly detection method based on digital twin technology according to claim 1 is characterized in that: The method further comprises: Based on the anomaly analysis results and real-time data stream within the first target time period, calling the digital twin model to simulate the operation of the target road section within a second target time period to obtain a first simulation result, wherein the second target time period is a future time period after the first target time period; Analyzing the simulated operation status recorded in the first simulation result based on the preset abnormal behavior rule to determine a probability value of an abnormal event occurring on the target road section within the second target time period; When the probability value is not zero, obtaining a second simulation result of the digital twin model, wherein the second simulation result includes: the probability value, the time prediction data, and the space prediction data; An abnormality prediction report for the target road section within the second target time period is generated based on the second simulation result, and an abnormality warning is triggered based on the abnormality prediction report.

3. The rail transit anomaly detection method based on digital twin technology according to claim 1 or 2, characterized in that: The method further comprises: Based on the position data or spatial prediction data determined by the digital twin model, marking the abnormal range on a map in the digital twin model, and / or marking the abnormal equipment point on a three-dimensional model in the digital twin model, wherein the rail transit scene is recorded in a two-dimensional form in the map and the three-dimensional model records the rail transit scene in a three-dimensional form; Display the map and / or three-dimensional model after the anomaly is marked.

4. The rail transit anomaly detection method based on digital twin technology according to claim 1 is characterized in that: After collecting environmental perception data and video surveillance data of at least one target road section, the method further includes: Performing data preprocessing on all collected data, wherein the data preprocessing includes at least the following operations: data cleaning operation and data fusion operation; The data cleaning operation includes at least one of the following: removing noise, removing duplicate data, and removing invalid data; The data fusion operation includes: performing time synchronization and spatial matching on the environmental perception data and the video surveillance data to form a consistent data set.

5. The rail transit anomaly detection method based on digital twin technology according to claim 4 is characterized in that: The data preprocessing also includes: data mining operations, The data mining operation includes: analyzing the consistent data set obtained after fusion to obtain a key feature set, wherein the key feature set includes abnormal behavior features and structural stress change features in the target road section.

6. The rail transit anomaly detection method based on digital twin technology according to claim 1 is characterized in that: The steps of constructing the digital twin model include: Using a geographic information system to obtain geographic information of all rail transit sections within the target range, and constructing a basic geographic framework of all rail transit sections within the target range based on the geographic information; Using a building information model to obtain infrastructure information of all rail transit sections within the target range, and constructing a three-dimensional equipment model of all infrastructure in all rail transit sections within the target range based on the infrastructure information; The basic geographic framework and the three-dimensional equipment model are integrated to obtain the digital twin models corresponding to all rail transit sections within the target range.

7. The rail transit anomaly detection method based on digital twin technology according to claim 6 is characterized in that: After obtaining the digital twin model, the following steps are also included: Acquire real-time data collected by all the infrastructure devices within the target range through Internet of Things technology; The operating status parameters of the corresponding virtual devices in the digital twin model are updated based on the real-time data.

8. The rail transit anomaly detection method based on digital twin technology according to claim 6 is characterized in that: The step of fusing the basic geographic framework and the three-dimensional device model includes: Extracting location information of each basic device recorded in the basic geographic framework, wherein the location information includes at least one of the following: coordinate data, detection direction, and detection coverage; Drawing a map of all rail transit sections within the target range based on the basic geographic framework, and calibrating the virtual position information of the corresponding virtual device based on the coordinate data of each basic device; A three-dimensional model of all rail transit sections within the target range is established based on the calibrated map, and a three-dimensional device model corresponding to each virtual device in the three-dimensional model is calibrated based on the detection direction and the detection coverage range.

9. A rail transit anomaly detection device based on digital twin technology, characterized in that: include: a collection unit, configured to collect environmental perception data and video surveillance data of at least one target road section to obtain a real-time data stream within a first target time period; an abnormality analysis unit, configured to perform an abnormality analysis on the real-time data stream of the target road section within the first target time period based on preset abnormal behavior rules to obtain an abnormality analysis result; a detection unit, configured to call a digital twin model and detect the digital twin model based on the anomaly analysis result and the real-time data stream to obtain a detection result, wherein the digital twin model is a pre-built virtual model for synchronously reflecting the actual physical environment, basic equipment layout, and equipment operating status in the rail transit scene, and the detection result is used to record the location data and impact range of the abnormal event; A warning unit is configured to generate an abnormality detection report for the target road section within the first target time period based on the location data and impact range of the abnormal event, and trigger an abnormality warning based on the abnormality detection report.

10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the rail transit anomaly detection method based on digital twin technology as described in any one of claims 1 to 7.

Images