A Rapid Response Method and System for Traffic Incidents Based on Multi-Source Data Fusion

Abstract

This application provides a method and system for rapid response to traffic incidents based on multi-source data fusion, relating to the field of traffic incident response technology. The method includes: extracting traffic feature vectors from multi-source traffic data for multi-source fusion calculation; determining whether abnormal traffic incidents exist in the fused situational data; when abnormal traffic incidents exist, determining the type and locating the information; generating an initial response plan based on the determined type and location information; pushing the initial response plan to the handling department's terminal and receiving execution status information from the handling department; recording the timestamps and operation logs of the entire process, and evaluating the effectiveness of this response process. This application addresses the technical problem of lacking multi-source traffic data fusion analysis in existing technologies, enabling rapid response to traffic incidents based on multi-source traffic data fusion, and achieving the technical effect of improving the accuracy of traffic incident identification.

Description

Technical Field

[0001] This application relates to the field of traffic incident response technology, and in particular to a rapid response method and system for traffic incidents based on multi-source data fusion. Background Technology

[0002] With the continuous growth of urban motor vehicle ownership and the increasing complexity of urban road traffic network structure, the traffic operation environment is gradually showing the characteristics of high density, high dynamics and multi-factor coupling. Various traffic incidents such as traffic accidents, vehicle malfunctions, sudden congestion and severe weather occur frequently during road operation. Once a traffic incident occurs, it often quickly has a chain reaction on surrounding road sections and even the entire regional road network. Therefore, traffic management departments have put forward higher requirements for the rapid identification and timely handling of traffic incidents.

[0003] Currently, existing traffic incident response systems typically rely on a single type of data source for traffic status assessment. For example, they may depend solely on traffic flow detection equipment to obtain parameters such as vehicle volume and speed, or rely on video surveillance systems for manual or semi-automatic incident identification. In actual operation, different types of data are often collected and managed independently by different systems, lacking a unified data fusion and collaborative analysis mechanism. This results in traffic operation status information exhibiting problems such as limited data dimensions, low information correlation, and insufficient overall situational expression. When a traffic incident occurs, the traffic status reflected by a single data source often fails to comprehensively depict the actual traffic situation. For instance, traffic flow parameters can only reflect the macroscopic traffic conditions and cannot accurately identify specific incident types. While video image information can reflect the scene, it is easily affected by lighting, occlusion, and weather conditions in complex environments, thus reducing identification accuracy. Without the collaborative fusion of multi-source information, traffic management systems struggle to form a complete understanding of the traffic operation situation, thereby affecting the reliability of traffic incident identification results.

[0004] In summary, the existing technology suffers from a lack of technical mechanisms for unified fusion and analysis of multi-source traffic data, such as traffic flow parameters, video images, vehicle trajectories, and environmental meteorological data. This makes it difficult for traffic management systems to accurately construct comprehensive traffic situation information that reflects the road's operational status, further affecting the accuracy and timeliness of traffic incident identification and response decisions. Summary of the Invention

[0005] The purpose of this application is to provide a rapid response method and system for traffic incidents based on multi-source data fusion, in order to solve the technical problem in the prior art that the lack of a technical mechanism for unified fusion and analysis of multi-source traffic data such as traffic flow parameters, video images, vehicle trajectories and environmental meteorology makes it difficult for traffic management systems to accurately construct comprehensive traffic situation information reflecting the road operating status, which further affects the accuracy and timeliness of traffic incident identification and response decisions.

[0006] In view of the above problems, this application provides a method and system for rapid response to traffic incidents based on multi-source data fusion.

[0007] Firstly, this application provides a rapid response method for traffic incidents based on multi-source data fusion, implemented through a rapid response system for traffic incidents based on multi-source data fusion. The method includes: collecting multi-source traffic data containing traffic flow parameters, video images, vehicle trajectory data, and environmental meteorological information; extracting traffic feature vectors reflecting traffic operation status from the multi-source traffic data, and performing multi-source fusion calculations on the traffic feature vectors using a data fusion algorithm to generate fused situational data; inputting the fused situational data into a pre-constructed traffic incident identification model for analysis to determine whether an abnormal traffic incident exists; when an abnormal traffic incident exists, combining the fused situational data to determine the type and location of the abnormal traffic incident; querying an emergency response plan database based on the determined type and location information to generate an initial response plan including resource scheduling and traffic management strategies; pushing the initial response plan to the handling department's terminal and receiving execution status information from the handling department to construct a closed-loop process for incident handling; and recording the timestamps and operation logs of the entire process after incident handling is completed, and evaluating the effectiveness of the response process.

[0008] Preferably, the rapid response method for traffic incidents based on multi-source data fusion further includes: collecting traffic flow parameters such as vehicle volume, speed, and occupancy rate at a preset frequency using geomagnetic sensors and radar sensors; collecting video images covering the target road segment using surveillance cameras; acquiring vehicle trajectory data of floating cars using in-vehicle navigation terminals and mobile communication networks; collecting environmental meteorological information including rainfall, visibility, and road surface temperature using meteorological monitoring stations deployed along the road; and combining the traffic flow parameters, the video images, the vehicle trajectory data, and the environmental meteorological information to obtain the multi-source traffic data.

[0009] Preferably, the rapid response method for traffic incidents based on multi-source data fusion further includes: detecting missing values ​​in the multi-source traffic data and interpolating and filling in missing values ​​using historical data from adjacent time periods; filtering noise interference in the interpolated and filled data and aligning the filtered data with timestamps; performing coordinate transformation and projection transformation on the timestamp-aligned data to unify it to a preset geographic coordinate system; and extracting and encoding keyframes from the unified coordinate system data to generate preprocessed multi-source traffic data.

[0010] Preferably, the rapid response method for traffic incidents based on multi-source data fusion further includes: extracting traffic feature vectors of average vehicle speed, traffic density, headway, and queue length from the multi-source traffic data; calculating the credibility weight value of the traffic feature vectors for the same traffic feature; performing fusion calculation on the traffic feature vectors using a weighted average algorithm based on the credibility weight value; mapping the fused traffic feature vectors onto the road network topology and outputting the fused situational data.

[0011] Preferably, the rapid response method for traffic incidents based on multi-source data fusion further includes: using a trained deep neural network model to abstract and extract spatiotemporal features from the fused situational data layer by layer; inputting the extracted high-level features into a classifier to calculate the probability value that the current state belongs to an abnormal traffic incident; setting a judgment threshold and comparing the probability value with the judgment threshold; if the probability value exceeds the judgment threshold, outputting a judgment result that an abnormal traffic incident exists.

[0012] Preferably, the rapid response method for traffic incidents based on multi-source data fusion further includes: calling video images and using computer vision algorithms to perform secondary analysis and identification of the incident scene of the abnormal traffic incident; determining the judgment type based on the visual recognition results; calculating the offset of the incident point relative to the reference point by combining the distance data measured by the radar sensor and the calibration parameters of the camera; superimposing the offset onto the geographic coordinates of the camera to calculate the latitude and longitude coordinates of the incident point to obtain the positioning information.

[0013] Preferably, the rapid response method for traffic incidents based on multi-source data fusion further includes: matching the corresponding emergency response process template according to the determination type; obtaining the real-time distribution of police resources, rescue vehicles, and variable message signs around the location information to obtain the road network traffic flow status; calculating the optimal driving route for rescue vehicles to reach the incident scene based on the road network traffic flow status; generating diversion guidance strategies for surrounding vehicles and information release content for the information signs; and combining the response process template, resource scheduling path, and guidance strategy to generate an initial response plan.

[0014] Secondly, this application also provides a rapid traffic incident response system based on multi-source data fusion, used to execute the rapid traffic incident response method based on multi-source data fusion as described in the first aspect, comprising: a multi-source traffic data acquisition module, used to acquire multi-source traffic data including traffic flow parameters, video images, vehicle trajectory data, and environmental meteorological information; a fusion situation data generation module, used to extract traffic feature vectors reflecting traffic operation status from the multi-source traffic data, and use a data fusion algorithm to perform multi-source fusion calculation on the traffic feature vectors to generate fusion situation data; and an anomaly judgment module, used to input the fusion situation data into a pre-constructed traffic incident identification model for analysis and judgment. The system includes: a type determination module for determining the type and location of an abnormal traffic event based on the fused situational data; an initial response plan generation module for querying the emergency response plan database based on the determined type and location information to generate an initial response plan that includes resource scheduling and traffic management strategies; a closed-loop process construction module for pushing the initial response plan to the handling department's terminal and receiving execution status information from the handling department to construct a closed-loop process for event handling; and an effectiveness evaluation module for recording the timestamps and operation logs of the entire process after the event handling is completed and evaluating the effectiveness of the response process.

[0015] The technical solution provided in this application has at least the following technical effects or advantages: by realizing rapid response to traffic incidents based on multi-source traffic data fusion, it can obtain the road network operation status in real time and comprehensively and accurately identify abnormal traffic incidents, thereby achieving the technical effects of improving the accuracy of traffic incident identification, shortening response time and optimizing the execution effect of emergency response plans.

[0016] The above description is merely an overview of the technical solution of this application. To enable a clearer understanding of the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the rapid response method for traffic incidents based on multi-source data fusion proposed in this application.

[0019] Figure 2 This is a schematic diagram of the structure of the traffic incident rapid response system based on multi-source data fusion proposed in this application.

[0020] Figure labeling: 1. Multi-source traffic data acquisition module; 2. Fusion situation data generation module; 3. Anomaly detection module; 4. Type detection module; 5. Initial response plan generation module; 6. Closed-loop process construction module; 7. Performance evaluation module. Detailed Implementation

[0021] This application provides a method and system for rapid traffic incident response based on multi-source data fusion. It addresses the technical problem in existing technologies where the lack of a unified fusion and analysis mechanism for multi-source traffic data such as traffic flow parameters, video images, vehicle trajectories, and environmental meteorological data makes it difficult for traffic management systems to accurately construct comprehensive traffic situation information reflecting road operating conditions, further affecting the accuracy and timeliness of traffic incident identification and response decisions. The method achieves rapid traffic incident response based on multi-source traffic data fusion, enabling real-time and comprehensive acquisition of road network operating conditions and accurate identification of abnormal traffic incidents. This results in improved accuracy of traffic incident identification, shortened response time, and optimized execution of emergency response plans.

[0022] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0023] Example 1, please refer to the appendix. Figure 1 This application provides a rapid response method for traffic incidents based on multi-source data fusion, which is applied to a rapid response system for traffic incidents based on multi-source data fusion. The method specifically includes the following steps:

[0024] S1: Collect multi-source traffic data including traffic flow parameters, video images, vehicle trajectory data, and environmental meteorological information.

[0025] Furthermore, this application also includes: collecting traffic flow parameters such as vehicle volume, vehicle speed, and occupancy rate at a preset frequency using geomagnetic sensors and radar sensors; collecting video images covering the target road section using surveillance cameras; obtaining vehicle trajectory data of floating cars using in-vehicle navigation terminals and mobile communication networks; collecting environmental meteorological information including rainfall, visibility, and road surface temperature using meteorological monitoring stations deployed along the road; and combining the traffic flow parameters, the video images, the vehicle trajectory data, and the environmental meteorological information to obtain the multi-source traffic data.

[0026] Furthermore, this application also includes: performing missing value detection on the multi-source traffic data and interpolating and filling it using historical data from adjacent time periods; filtering the noise interference present in the interpolated and filled data and aligning the filtered data with timestamps; performing coordinate transformation and projection transformation on the timestamp-aligned data to unify it to a preset geographic coordinate system; and extracting and encoding keyframes from the unified coordinate system data to generate preprocessed multi-source traffic data.

[0027] Specifically, traffic flow parameters such as vehicle volume, vehicle speed, and occupancy are collected at preset frequencies using geomagnetic sensors and radar sensors. This can be understood as deploying traffic sensing devices such as geomagnetic sensors and radar sensors on road sections or lane areas. By setting a fixed time interval for collection, the operating status of vehicles passing through the road section is continuously detected. Among them, vehicle volume is used to represent the number of vehicles passing through the detection section per unit time, vehicle speed is used to represent the average speed of vehicles on the road, and occupancy rate is used to represent the degree of traffic congestion reflected by the proportion of time that vehicles occupy the detection area. Through the above multiple traffic operation indicators, a set of traffic flow parameter data that can characterize the road traffic operation status is formed, thereby providing a basic data source for traffic status analysis.

[0028] Furthermore, by using surveillance cameras to collect video images covering the target road segment, it can be understood as installing surveillance camera equipment with real-time video acquisition capabilities above or on both sides of the target road segment to continuously collect images of the road area. By acquiring continuous video image data of vehicle operation, lane occupancy, and the occurrence of emergencies within the road area, the traffic management system can obtain traffic operation information from a visual dimension, thereby providing intuitive image data support for traffic event recognition, vehicle behavior analysis, and traffic anomaly detection.

[0029] Furthermore, by acquiring vehicle trajectory data of floating cars through in-vehicle navigation terminals and mobile communication networks, it can be understood that the navigation terminal equipment installed in the vehicle continuously records information such as the vehicle's position coordinates, travel time, and travel speed in the road network, and uploads the recorded position data to the traffic information platform through the mobile communication network. This forms a trajectory data set that reflects the actual travel path of the vehicle in the road network. By aggregating and analyzing the trajectory information of multiple floating cars, the spatial distribution of traffic flow in the road network and the dynamic changes in vehicle operation can be reflected.

[0030] Furthermore, by deploying meteorological monitoring stations along the road, environmental meteorological information including rainfall, visibility, and road surface temperature is collected. This can be understood as deploying meteorological monitoring equipment along the road to monitor the road operating environment in real time. Rainfall is used to characterize the impact of rainfall intensity on road slipperiness, visibility is used to characterize the impact of the atmospheric environment on driving visibility conditions, and road surface temperature is used to characterize the impact of road surface thermal state on vehicle braking performance and driving safety. By continuously collecting the above meteorological indicators, an environmental meteorological information data set that can reflect changes in the road environment is formed.

[0031] Furthermore, by combining the traffic flow parameters, the video images, the vehicle trajectory data, and the environmental meteorological information, the multi-source traffic data is obtained. This can be understood as the unified collection and integration of various types of data obtained from traffic detection equipment, video surveillance equipment, vehicle terminal equipment, and meteorological monitoring equipment. By uniformly managing and organizing data from different sources in both time and spatial dimensions, traffic operation information from different sensing channels is formed into a data set with multi-dimensional attributes, thereby constructing a multi-source traffic data foundation system that includes traffic operation status, vehicle behavior characteristics, and road environment status.

[0032] Missing value detection is performed on the multi-source traffic data, and interpolation is performed using historical data from adjacent time periods. This can be understood as performing integrity analysis on the data set acquired by traffic flow detection equipment, video acquisition equipment, vehicle terminal equipment, and environmental monitoring equipment. By scanning and identifying each item of time series data, the location of missing information in the data records is determined. Missing value detection is used to identify data items that were not collected or failed to be transmitted at a specific time node, while interpolation is used to estimate the value based on historical data recorded in adjacent time periods before and after the missing time node. Data values ​​for the corresponding time positions are generated through mathematical interpolation methods, thereby ensuring the continuity and integrity of traffic data in the time dimension.

[0033] Furthermore, filtering out noise interference in the interpolated data and aligning the timestamps on the filtered data can be understood as optimizing the quality of the interpolated dataset. Noise interference represents random fluctuations or abnormal deviations that occur during data acquisition, transmission, or equipment operation. The filtering algorithm smooths the data sequence to reduce the impact of random interference on the data analysis results. Timestamp alignment is used to uniformly adjust the recording times from different data sources. By synchronizing the timestamps generated by different devices, different data sources can establish a correspondence on the same time dimension, thereby achieving a consistent expression of multi-source traffic data in the time dimension.

[0034] Furthermore, performing coordinate transformation and projection transformation on timestamp-aligned data to unify it to a preset geographic coordinate system can be understood as performing spatial coordinate standardization processing on data containing spatial location information. Coordinate transformation is used to convert the original coordinate representation methods used by different devices, such as converting pixel coordinates, local planar coordinates, or device relative coordinates into a unified geographic location expression method. Projection transformation is used to map the coordinates of the Earth's surface to a planar map coordinate system. By pre-setting a unified geographic coordinate system, data from different sources can be represented under the same spatial reference frame, thereby achieving a unified expression of traffic data in the spatial dimension.

[0035] Furthermore, keyframe extraction and encoding transformation are performed on the unified coordinate system data to generate preprocessed multi-source traffic data. This can be understood as information compression and structuring processing of a spatially unified dataset. Keyframe extraction is used to select typical data frames that can represent traffic state change characteristics from continuous video images or continuous data sequences. By retaining representative time node information, the redundancy of data is reduced. Encoding transformation is used to convert and store different types of data according to a unified data structure or data format, so that traffic flow parameter data, video image feature data, and vehicle trajectory data can be managed in a unified data expression form, thereby generating a preprocessed dataset that can be used by subsequent traffic state analysis and event recognition models.

[0036] S2: Extract traffic feature vectors reflecting traffic operation status from the multi-source traffic data, and use a data fusion algorithm to perform multi-source fusion calculation on the traffic feature vectors to generate fused situation data.

[0037] Furthermore, this application also includes: extracting traffic feature vectors of average vehicle speed, traffic density, headway, and queue length from the multi-source traffic data; calculating the confidence weight value of the traffic feature vectors for the same traffic feature; performing a weighted average algorithm to fuse the traffic feature vectors according to the confidence weight value; mapping the fused traffic feature vectors onto the road network topology and outputting the fused situational data.

[0038] Specifically, extracting traffic feature vectors such as average vehicle speed, traffic density, headway, and queue length from the multi-source traffic data can be understood as performing feature extraction processing on the traffic operation information contained in the multi-source traffic data after data preprocessing. By analyzing the changes in vehicle operation status and road traffic flow, key indicators that can reflect traffic operation characteristics are calculated. Among them, average vehicle speed is used to represent the average speed of vehicles on a road segment per unit time, traffic density is used to represent the number of vehicles contained within a unit length of road, headway is used to represent the time interval between two adjacent vehicles passing the same detection position, and queue length is used to represent the length of the queue space of vehicles on the road under traffic congestion or signal control conditions. By combining and expressing the above multiple traffic operation indicators in vector form, a traffic feature vector that can comprehensively reflect changes in traffic operation status is formed.

[0039] Furthermore, calculating the credibility weight value of traffic feature vectors for the same traffic feature can be understood as performing reliability assessment processing on traffic feature vectors from different data sources. The credibility weight value is used to represent the degree of data credibility of different data sources when reflecting the same traffic feature. By comprehensively analyzing the collection accuracy, data stability and data integrity of each data source, the contribution of different traffic feature vectors in the fusion calculation process is determined, thereby providing a basis for weight parameters for subsequent multi-source data fusion.

[0040] Furthermore, based on the credibility weight values, a weighted average algorithm is used to fuse the traffic feature vectors. This can be understood as follows: after obtaining the weight parameters corresponding to each traffic feature vector, the traffic feature data from multiple sources are comprehensively processed using a weighted average calculation method. The weighted average algorithm is used to perform numerical calculations on each traffic feature vector according to different weight ratios, so that data with higher credibility occupies a higher proportion in the fusion result, thereby generating a unified traffic feature expression result that can comprehensively reflect multi-source traffic information.

[0041] Furthermore, the fused traffic feature vectors are mapped onto the road network topology to output the fused situational data. This can be understood as associating the traffic feature results obtained through fusion calculation with the road network structure model. The road network topology is a structured model representing road nodes, road connection relationships, and the spatial distribution of roads. By spatially mapping the fused traffic feature vectors according to corresponding road segments or road network nodes, traffic operation status information can be visualized and structured in the road network structure, thereby forming fused situational data describing the overall traffic operation status of the road network.

[0042] S3: Input the fused situational data into the pre-built traffic event identification model for analysis to determine whether there are any abnormal traffic events.

[0043] Furthermore, this application also includes: using a trained deep neural network model to abstract and extract spatiotemporal features from the fused situational data layer by layer; inputting the extracted high-level features into a classifier to calculate the probability value of the current state belonging to an abnormal traffic event; setting a judgment threshold and comparing the probability value with the judgment threshold; if the probability value exceeds the judgment threshold, outputting a judgment result indicating the existence of an abnormal traffic event.

[0044] Specifically, the use of a trained deep neural network model to abstract and extract spatiotemporal features from the fused situational data can be understood as follows: after completing the fusion of multi-source data and obtaining fused situational data that reflects the road traffic operation status, the fused situational data is input into a deep neural network model trained with sample data for feature learning processing. The deep neural network model is used to construct a nonlinear mapping relationship with a multi-layer structure. Through the multi-layer neural network structure, the input data is processed layer by layer, thereby realizing the automatic learning and expression of traffic data features. Spatiotemporal features are used to represent the changing trend of traffic operation status in the time dimension and the distribution relationship in the spatial dimension. Through the layer-by-layer abstraction and extraction process, the original traffic status data is gradually transformed into a high-level semantic feature representation that reflects the traffic operation pattern.

[0045] Furthermore, the extracted high-level features are input into the classifier to calculate the probability value of the current state belonging to an abnormal traffic event. This can be understood as passing the high-level feature representation output by the deep neural network model as input data to the classification decision module for analysis and processing. The high-level features are used to represent the traffic state feature information after being abstracted and expressed by the deep learning model, while the classifier is used to determine the category of the input features according to the established classification model. By calculating the probability of the traffic state features, the probability value of the current traffic operation state belonging to the abnormal traffic event category is obtained, thereby realizing a quantitative assessment of the possibility of abnormal traffic conditions.

[0046] Furthermore, setting a judgment threshold and comparing the probability value with the judgment threshold can be understood as pre-setting a probability limit parameter for judging abnormal states in the process of traffic event recognition. The judgment threshold is used to represent the minimum probability standard for the recognition model output result to be identified as an abnormal traffic event. By comparing the abnormal probability value calculated by the classifier with the preset threshold, it is determined whether the current traffic operation state meets the judgment conditions for abnormal event recognition, thereby realizing a standardized judgment mechanism in the process of traffic abnormal state recognition.

[0047] Furthermore, if the probability value exceeds the judgment threshold, the judgment result of the existence of an abnormal traffic event is output. This can be understood as the traffic event identification module generating the corresponding abnormal event identification result when the probability comparison result meets the abnormal judgment condition. The abnormal traffic event is used to represent the abnormal state of traffic operation caused by factors such as traffic accidents, vehicle failures, road congestion, or sudden environmental changes. By outputting the judgment result of the abnormal traffic event, the traffic management system can promptly identify abnormal traffic conditions that may occur in the road network, thereby providing a basis for subsequent event type analysis and emergency response decision-making.

[0048] S4: When the abnormal traffic event exists, the abnormal traffic event is determined and located by combining the fused situational data.

[0049] Furthermore, this application also includes: calling video images and using computer vision algorithms to perform secondary analysis and identification of the event scene of the abnormal traffic event; determining the judgment type based on the visual recognition results; calculating the offset of the event occurrence point relative to the reference point by combining the distance data measured by the radar sensor and the calibration parameters of the camera; superimposing the offset onto the geographic coordinates of the camera to calculate the latitude and longitude coordinates of the event occurrence point to obtain the positioning information.

[0050] Specifically, calling video images and using computer vision algorithms to perform secondary analysis and identification of the event scene of the abnormal traffic event can be understood as follows: after the traffic event identification model detects an abnormal traffic state, video image data corresponding to the abnormal occurrence area is retrieved from the road monitoring system. By further analyzing and processing the road scene contained in the monitoring video, the accuracy of event identification is improved. Here, video images are used to represent a continuous visual data sequence collected by road monitoring camera equipment, and computer vision algorithms are used to automatically analyze and extract features from image or video data. Through methods such as target detection, target tracking, and scene understanding, the behavioral states of vehicles, pedestrians, and obstacles are identified, thereby enabling a more detailed identification and analysis of the event scene corresponding to the abnormal traffic state.

[0051] Furthermore, determining the judgment type based on the visual recognition results can be understood as performing category judgment processing on the recognition information output by the visual recognition module after completing the image analysis of the event scene. The visual recognition results are used to represent scene feature information such as vehicle collisions, vehicle stagnation, road obstacle appearances, or traffic congestion obtained through computer vision algorithms, while the judgment type is used to clearly identify the category to which the abnormal traffic event belongs. By performing rule matching or model classification analysis on the visual recognition results, the abnormal traffic state is divided into specific event categories so that the subsequent emergency response process can take corresponding response measures according to different event types.

[0052] Furthermore, by combining the distance data measured by the radar sensor with the calibration parameters of the camera, the offset of the event occurrence point relative to the reference point is calculated. This can be understood as calculating the precise spatial location of the abnormal event after completing the event type identification. The radar sensor is used to measure the distance of the target object in the road area through the principle of electromagnetic wave reflection. The distance data is used to represent the spatial distance relationship between the target object and the sensor. The camera calibration parameters are used to describe the intrinsic and extrinsic parameters determined during the installation of the camera equipment. By using the imaging geometry of the camera equipment and the radar ranging results for joint calculation, the spatial location of the abnormal event target in the monitoring field of view can be determined, and the spatial offset of the event occurrence point relative to the preset reference point can be further calculated.

[0053] Furthermore, the offset is superimposed on the geographic coordinates of the camera to calculate the latitude and longitude coordinates of the event location, thus obtaining the positioning information. This can be understood as follows: after obtaining the spatial offset of the abnormal event relative to the monitoring device, the offset is superimposed on the geographic location of the camera device on the map. The geographic coordinates of the camera are used to represent the longitude and latitude position of the monitoring device in the geographic information system. The relative position offset is mapped to the geographic coordinate system through a spatial coordinate transformation method, thereby calculating the latitude and longitude coordinate information corresponding to the location of the abnormal event. The latitude and longitude coordinates are used to accurately identify the location of the traffic event in the geographic information system, thereby forming positioning information for traffic management and emergency dispatch.

[0054] S5: Based on the determination type and location information, query the emergency response plan database and generate an initial response plan that includes resource scheduling and diversion strategies.

[0055] Furthermore, this application also includes: matching the corresponding emergency response process template according to the determination type; obtaining the real-time distribution of police resources, rescue vehicles and variable message signs around the location information to obtain the road network traffic flow status; calculating the optimal driving route for rescue vehicles to reach the incident scene based on the road network traffic flow status; generating diversion guidance strategies for surrounding vehicles and information release content for information signs; and combining the response process template, resource scheduling path and guidance strategy to generate an initial response plan.

[0056] Specifically, based on the determination type, the corresponding emergency response process template is matched. This can be understood as follows: after identifying the type of abnormal traffic incident, the traffic management system accesses a pre-established emergency plan database and retrieves the response process structure corresponding to the identified incident category from the database. The determination type indicates the specific category to which the traffic incident belongs, such as traffic accident, vehicle breakdown, road blockage, or traffic congestion. The emergency response process template describes the standardized handling steps and operating procedures formulated for different incident categories, including information reporting procedures, on-site handling steps, resource dispatch sequence, and traffic diversion measures. Through the matching process, the incident type is associated with the corresponding response process, thereby forming a response process framework applicable to the current traffic incident.

[0057] Furthermore, by acquiring the real-time distribution of police resources, rescue vehicles, and variable message signs around the location information, the road network traffic flow status can be obtained. This can be understood as acquiring real-time information on traffic management resources and road operation status around the incident area after determining the handling process. The location information is used to represent the geographic coordinate data corresponding to the location of the abnormal traffic incident. The police resources are used to represent the on-duty personnel and patrol vehicle resources participating in traffic control tasks in the road traffic management system. The rescue vehicles are used to represent the professional vehicles and equipment that can participate in traffic accident rescue or road clearing tasks. The variable message signs are used to represent the electronic display devices set up along the road to publish traffic information prompts. By acquiring the real-time spatial distribution of the above resources in the road network and combining it with road vehicle operation data for comprehensive analysis, road network traffic flow status information reflecting the current road network traffic operation status can be formed.

[0058] Furthermore, based on the road network traffic flow status, the optimal driving route for rescue vehicles to reach the incident site is calculated. This can be understood as, after understanding the road traffic operation status, using a path planning algorithm to calculate and analyze the driving route of the rescue vehicle from its current location to the location of the abnormal event. The road network traffic flow status describes the vehicle density, traffic speed, and congestion level of each road segment in the road network, while the optimal driving route represents the optimal route determined under the comprehensive consideration of factors such as road traffic efficiency, distance, and traffic congestion level. Through path optimization calculation, rescue vehicles can reach the incident site in a shorter time, thereby improving the efficiency of emergency response to traffic incidents.

[0059] Furthermore, generating diversion guidance strategies for surrounding vehicles and information dissemination content on information boards can be understood as guiding and controlling traffic flow around the event-affected area after the rescue route planning is completed. Traffic management systems formulate traffic guidance strategies for surrounding vehicles. The diversion guidance strategy is used to guide some vehicles to avoid the event-affected area by adjusting their routes, thereby reducing road congestion. The information dissemination content on information boards is used to release real-time traffic information to drivers through variable message signs along the road, such as detour suggestions, congestion warnings, and road control information. The dissemination of traffic guidance information can guide vehicles to divert reasonably, thereby mitigating the impact of abnormal events on the overall traffic operation.

[0060] Furthermore, combining the handling process template, resource dispatch path, and guidance strategy to generate an initial response plan can be understood as integrating and processing various information after completing the handling process matching, resource status acquisition, path planning, and traffic guidance strategy generation. The handling process template is used to guide the specific operational steps in the traffic incident handling process, the resource dispatch path is used to indicate the driving routes of police resources and rescue vehicles to the incident scene, and the guidance strategy is used to guide traffic vehicles to be diverted reasonably. By unifying and integrating the above multiple handling information, a comprehensive handling plan that includes incident handling steps, resource dispatch arrangements, and traffic diversion measures can be formed, thereby obtaining an initial response plan to guide traffic management departments in carrying out emergency response work.

[0061] S6: Push the initial response plan to the handling department terminal and receive the execution status information fed back by the handling department to build a closed-loop process for event handling.

[0062] Specifically, the initial response plan is pushed to the handling department's terminal, and the execution status information fed back by the handling department is received, constructing a closed-loop process for incident handling. This can be understood as follows: after forming a comprehensive handling plan for abnormal traffic incidents, the generated initial response plan is sent through the traffic management information system to the information terminal equipment used by the relevant execution agency responsible for handling traffic incidents. The initial response plan represents a comprehensive set of handling instructions including emergency response procedures, resource allocation arrangements, and traffic diversion strategies. The handling department's terminal represents the mobile communication terminal or command and dispatch terminal equipment used by the traffic management department, public security traffic police department, or emergency rescue agency. The system enables real-time transmission of response plans through a data communication network, allowing relevant personnel to obtain traffic incident handling tasks and corresponding execution steps. Simultaneously, during the traffic incident handling process, the handling department's terminal continuously transmits execution status information back to the traffic management system. This execution status information indicates the progress of various tasks, on-site handling conditions, and resource usage status during the traffic incident handling process. By receiving and recording the above feedback data, the traffic management system can continuously monitor the progress of incident handling and dynamically update the handling process based on the feedback results, thereby forming a complete closed-loop process for incident handling that includes the issuance of response plans, on-site execution, and the transmission of execution feedback information.

[0063] S7: After the incident is resolved, record the timestamps and operation log data of the entire process, and evaluate the effectiveness of the response process.

[0064] Specifically, after the incident is handled, timestamps and operation logs of the entire process are recorded, and the effectiveness of the response is evaluated. This can be understood as the systematic recording and analysis of relevant operational information throughout the entire incident response process by the traffic management system after the abnormal traffic incident handling procedure is completed. "Incident handling completed" indicates that tasks such as clearing traffic accidents, restoring road access, or alleviating traffic congestion have been completed and traffic has returned to normal. Timestamps represent the time records corresponding to each key operational node in the incident response process, including incident identification time, response plan generation time, handling instruction issuance time, and on-site handling completion time. Recording the operation time at each stage forms a complete time series. According to reports, operation log data is used to record the specific operational behaviors performed by system modules and personnel during traffic incident response, such as incident identification, resource scheduling, route planning calculation, and traffic guidance information dissemination. Systematically storing these operational behaviors can form a complete record of the incident handling process. After data recording is completed, the traffic management system comprehensively analyzes the collected timestamp data and operation log data to evaluate the overall operational effectiveness of the traffic incident response process. Among them, the efficiency evaluation is used to quantitatively evaluate the efficiency and quality of traffic incident handling by analyzing indicators such as response time, resource scheduling efficiency, and traffic recovery speed, thereby providing data basis for subsequent optimization of traffic incident response strategies.

[0065] In summary, the traffic incident rapid response method based on multi-source data fusion provided in this application has the following technical effects: by realizing rapid response to traffic incidents based on multi-source traffic data fusion, it is possible to obtain the road network operation status in real time and comprehensively and accurately identify abnormal traffic incidents, thereby achieving the technical effects of improving the accuracy of traffic incident identification, shortening response time, and optimizing the implementation effect of emergency response plans.

[0066] Example 2: Based on the same inventive concept as the traffic incident rapid response method based on multi-source data fusion in the foregoing examples, this application also provides a traffic incident rapid response system based on multi-source data fusion. Please refer to the appendix. Figure 2The system includes: a multi-source traffic data acquisition module 1, used to collect multi-source traffic data including traffic flow parameters, video images, vehicle trajectory data, and environmental meteorological information; a fusion situation data generation module 2, used to extract traffic feature vectors reflecting traffic operation status from the multi-source traffic data, and use a data fusion algorithm to perform multi-source fusion calculation on the traffic feature vectors to generate fusion situation data; an anomaly judgment module 3, used to input the fusion situation data into a pre-built traffic event identification model for analysis to determine whether there is an abnormal traffic event; a type judgment module 4, used to determine the type and locate the abnormal traffic event by combining the fusion situation data when the abnormal traffic event exists; an initial response plan generation module 5, used to query the emergency response plan library according to the judgment type and location information to generate an initial response plan including resource scheduling and diversion strategies; a closed-loop process construction module 6, used to push the initial response plan to the terminal of the handling department, and receive the execution status information fed back by the handling department to construct a closed-loop process for event handling; and an effectiveness evaluation module 7, used to record the timestamps and operation log data of the entire process after the event handling is completed, and to evaluate the effectiveness of this response process.

[0067] Furthermore, the traffic incident rapid response system based on multi-source data fusion is also used to: collect traffic flow parameters such as vehicle volume, vehicle speed, and occupancy rate at a preset frequency using geomagnetic sensors and radar sensors; collect video images covering the target road section using surveillance cameras; obtain vehicle trajectory data of floating cars using in-vehicle navigation terminals and mobile communication networks; collect environmental meteorological information including rainfall, visibility, and road surface temperature using meteorological monitoring stations deployed along the road; and combine the traffic flow parameters, the video images, the vehicle trajectory data, and the environmental meteorological information to obtain the multi-source traffic data.

[0068] Furthermore, the traffic incident rapid response system based on multi-source data fusion is also used for: detecting missing values ​​in the multi-source traffic data and interpolating and filling in missing values ​​using historical data from adjacent time periods; filtering noise interference in the interpolated and filled data and aligning the filtered data with timestamps; performing coordinate transformation and projection transformation on the timestamp-aligned data to unify it to a preset geographic coordinate system; and extracting and encoding keyframes from the unified coordinate system data to generate preprocessed multi-source traffic data.

[0069] Furthermore, the traffic incident rapid response system based on multi-source data fusion is also used to: extract traffic feature vectors of average vehicle speed, traffic density, headway, and queue length from the multi-source traffic data; calculate the credibility weight value of the traffic feature vector for the same traffic feature; perform fusion calculation on the traffic feature vector using a weighted average algorithm according to the credibility weight value; map the fused traffic feature vector onto the road network topology structure, and output the fused situation data.

[0070] Furthermore, the traffic incident rapid response system based on multi-source data fusion is also used to: use a trained deep neural network model to abstract and extract spatiotemporal features from the fused situational data layer by layer; input the extracted high-level features into a classifier to calculate the probability value that the current state belongs to an abnormal traffic incident; set a judgment threshold and compare the probability value with the judgment threshold; if the probability value exceeds the judgment threshold, output the judgment result that there is an abnormal traffic incident.

[0071] Furthermore, the traffic incident rapid response system based on multi-source data fusion is also used to: call video images and use computer vision algorithms to perform secondary analysis and identification of the incident scene of the abnormal traffic incident; determine the judgment type based on the visual recognition results; calculate the offset of the incident point relative to the reference point by combining the distance data measured by the radar sensor and the calibration parameters of the camera; and superimpose the offset onto the geographic coordinates of the camera to calculate the latitude and longitude coordinates of the incident point to obtain the positioning information.

[0072] Furthermore, the traffic incident rapid response system based on multi-source data fusion is also used for: matching the corresponding emergency response process template according to the judgment type; obtaining the real-time distribution of police resources, rescue vehicles, and variable message signs around the location information to obtain the road network traffic flow status; calculating the optimal driving route for rescue vehicles to reach the incident scene based on the road network traffic flow status; generating diversion guidance strategies for surrounding vehicles and information release content for information signs; and combining the response process template, resource scheduling path, and guidance strategy to generate an initial response plan.

[0073] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The traffic incident rapid response method and specific examples based on multi-source data fusion in the foregoing embodiment one are also applicable to the traffic incident rapid response system based on multi-source data fusion in this embodiment. Through the foregoing detailed description of the traffic incident rapid response method based on multi-source data fusion, those skilled in the art can clearly understand the traffic incident rapid response system based on multi-source data fusion in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0074] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0075] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A rapid response method for traffic incidents based on multi-source data fusion, characterized in that, include: Collect multi-source traffic data including traffic flow parameters, video images, vehicle trajectory data, and environmental and meteorological information; Traffic feature vectors reflecting traffic operation status are extracted from the multi-source traffic data, and multi-source fusion calculations are performed on the traffic feature vectors using a data fusion algorithm to generate fused situation data. The fused situational data is input into a pre-built traffic incident identification model for analysis to determine whether any abnormal traffic incidents exist. When the abnormal traffic event exists, the abnormal traffic event is determined by type and information location by combining the fused situational data. Based on the determination type and location information, query the emergency response plan database to generate an initial response plan that includes resource scheduling and evacuation strategies; The initial response plan is pushed to the terminal of the handling department, and the execution status information fed back by the handling department is received to build a closed-loop process for event handling; After the incident is resolved, record the timestamps and operation logs of the entire process, and evaluate the effectiveness of the response process.

2. The rapid response method for traffic incidents based on multi-source data fusion as described in claim 1, characterized in that, Collect multi-source traffic data including traffic flow parameters, video images, vehicle trajectory data, and environmental meteorological information, including: Traffic flow parameters, including vehicle volume, speed, and occupancy, are collected at preset frequencies using geomagnetic sensors and radar sensors. Video images covering the target road section are collected using surveillance cameras; The vehicle trajectory data of the floating car is obtained through the vehicle navigation terminal and mobile communication network; By deploying meteorological monitoring stations along the road, environmental meteorological information including rainfall, visibility, and road surface temperature is collected; The multi-source traffic data is obtained by combining the traffic flow parameters, the video images, the vehicle trajectory data, and the environmental meteorological information.

3. The rapid response method for traffic incidents based on multi-source data fusion as described in claim 1, characterized in that, After collecting multi-source traffic data, the following is included: Missing values ​​are detected in the multi-source traffic data, and interpolation is performed using historical data from adjacent time periods to fill in the missing values. The noise interference in the interpolated data is filtered out, and the filtered data is timestamped. Perform coordinate transformation and projection transformation on the timestamp-aligned data to unify it to a preset geographic coordinate system; Keyframe extraction and encoding transformation are performed on data in a unified coordinate system to generate preprocessed multi-source traffic data.

4. The rapid response method for traffic incidents based on multi-source data fusion as described in claim 1, characterized in that, Traffic feature vectors reflecting traffic operation status are extracted from the multi-source traffic data, and multi-source fusion calculations are performed on the traffic feature vectors using a data fusion algorithm to generate fused situational data, including: Traffic feature vectors, including average vehicle speed, traffic density, headway, and queue length, are extracted from the multi-source traffic data. Calculate the confidence weight value of traffic feature vectors for the same traffic feature; Based on the confidence weight values, a weighted average algorithm is used to fuse and calculate the traffic feature vectors; The fused traffic feature vectors are mapped onto the road network topology, and the fused situational data is output.

5. The rapid response method for traffic incidents based on multi-source data fusion as described in claim 1, characterized in that, The fused situational data is input into a pre-built traffic event identification model for analysis to determine whether any abnormal traffic events exist, including: The spatiotemporal features in the fused situational data are extracted layer by layer using a trained deep neural network model. The extracted high-level features are input into the classifier to calculate the probability that the current state belongs to an abnormal traffic event. Set a judgment threshold, and compare the probability value with the judgment threshold; If the probability value exceeds the judgment threshold, the judgment result of the existence of an abnormal traffic event is output.

6. The rapid response method for traffic incidents based on multi-source data fusion as described in claim 1, characterized in that, When the aforementioned abnormal traffic event exists, the fused situational data is used to determine the type and locate the information of the abnormal traffic event, including: By calling up video images and using computer vision algorithms, a secondary analysis and identification of the scene of the abnormal traffic incident is performed. The judgment type is determined based on the visual recognition results; By combining the distance data measured by the radar sensor with the calibration parameters of the camera, the offset of the event occurrence point relative to the reference point is calculated. The offset is superimposed on the camera's geographic coordinates to calculate the latitude and longitude coordinates of the event location, thus obtaining the location information.

7. The rapid response method for traffic incidents based on multi-source data fusion as described in claim 1, characterized in that, Based on the assessment type and location information, the emergency response plan database is queried to generate an initial response plan that includes resource scheduling and evacuation strategies, including: Based on the determination type, match the corresponding emergency response procedure template; The system obtains the real-time distribution of police resources, rescue vehicles, and variable message signs in the vicinity of the location information to obtain the traffic flow status of the road network. Based on the road network traffic flow status, calculate the optimal driving route for rescue vehicles to reach the incident scene; Generate diversion and guidance strategies for surrounding vehicles and information content for information boards; The initial response plan is generated by combining the handling process template, resource scheduling path and guidance strategy.

8. A rapid response system for traffic incidents based on multi-source data fusion, characterized in that: The steps for implementing the rapid traffic incident response method based on multi-source data fusion as described in any one of claims 1 to 7 include: The multi-source traffic data acquisition module is used to collect multi-source traffic data, including traffic flow parameters, video images, vehicle trajectory data, and environmental meteorological information. The fusion situation data generation module is used to extract traffic feature vectors reflecting traffic operation status from the multi-source traffic data, and to perform multi-source fusion calculation on the traffic feature vectors using a data fusion algorithm to generate fusion situation data. The anomaly detection module is used to input the fused situational data into a pre-built traffic event identification model for analysis to determine whether there are any abnormal traffic events. The type determination module is used to determine the type and locate the abnormal traffic event by combining the fused situational data when the abnormal traffic event exists. The initial response plan generation module is used to query the emergency response plan library based on the judgment type and location information, and generate an initial response plan that includes resource scheduling and evacuation strategies. The closed-loop process construction module is used to push the initial response plan to the terminal of the handling department and receive the execution status information fed back by the handling department to construct a closed-loop process for event handling. The performance evaluation module is used to record the timestamps and operation log data of the entire process after the incident is handled, and to evaluate the performance of the response process.

Images