A highway traffic space-time service platform and method based on multi-source heterogeneous data fusion

Abstract

This invention provides a highway traffic spatiotemporal service platform and method based on multi-source heterogeneous data fusion. The platform includes: a data acquisition module; a data fusion management module; a base map creation module for generating a highway traffic spatiotemporal service base map with highway intersections as key nodes; a traffic situation analysis module for generating a traffic situation prediction map; and a decision support and personalized service module for providing real-time, visualized traffic situation displays, traffic guidance strategies, and personalized travel services to traffic managers and traffic participants based on the traffic situation prediction map. This invention not only provides real-time, visualized decision support for traffic managers but also allows for customized personalized travel plans for traffic participants, thereby achieving comprehensive perception and precise control of highway traffic operation status and greatly improving the level of intelligent traffic operation management.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation technology, specifically relating to a highway traffic spatiotemporal service platform and method based on multi-source heterogeneous data fusion. Background Technology

[0002] With the continuous improvement of the highway network, the efficiency and intelligence of highway traffic operation management have become crucial for the development of the transportation sector. However, with the sustained and rapid growth of traffic demand, traditional management models face increasingly severe challenges: frequent road congestion, accidents, extreme weather, and emergencies necessitate an urgent improvement in road network management capabilities. Existing highway traffic spatiotemporal service management models suffer from problems such as single data sources, insufficient information collection, limited analytical dimensions, and lagging decision support, making it difficult to accurately, in real-time, and personalizedly reflect the traffic operation status.

[0003] Specifically, existing systems still have significant gaps in research on the integration of geographic information data and spatiotemporal information services. Data at various levels of highway traffic are not fully integrated, lacking multi-source and dynamic integration of road, vehicle, and roadside information. In both temporal and spatial dimensions, data processing and comprehensive analysis are insufficient, failing to provide detailed characterization of indicators such as traffic flow, vehicle speed, and congestion index, thus hindering timely and accurate decision-making by managers. Furthermore, traditional platforms often operate in a decentralized manner, lacking a unified data sharing and management mechanism, making it impossible to achieve information interconnection and personalized service customization between traffic managers and users. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of the aforementioned background technology and provide a highway traffic spatiotemporal service platform and method based on multi-source heterogeneous data fusion. This platform not only provides real-time and visualized decision support for traffic managers but also allows for customized personalized travel plans for traffic participants, thereby achieving comprehensive perception and precise control of highway traffic operation status and greatly improving the level of intelligence in traffic operation management.

[0005] The technical solution adopted in this invention is: a highway traffic spatiotemporal service platform based on multi-source heterogeneous data fusion, comprising:

[0006] The data acquisition module is used to collect data including satellite remote sensing data, UAV imagery data, vehicle-mounted BeiDou GNSS trajectory data, and highway traffic data collected by roadside equipment.

[0007] The data fusion management module is used to construct a hierarchical highway traffic spatiotemporal service data warehouse by preprocessing, matching, integrating and representing the collected data, so as to realize the unified management of multi-source heterogeneous data.

[0008] The base map creation module is used to extract road network information based on the data warehouse using deep learning, edge detection and trajectory analysis methods, and to perform weighted fusion based on the weights corresponding to each extraction result to generate a spatiotemporal service base map of highway traffic with highway intersections as key nodes.

[0009] The traffic situation analysis module is used to perform time-series hierarchical analysis and convolutional neural network prediction of traffic flow, vehicle speed and congestion index in the data warehouse. Based on the spatiotemporal service base map of highway traffic, it combines cluster analysis with the radiation area based on highway intersections to detect abnormal traffic events and form a traffic situation prediction map.

[0010] The decision support and personalized service module is used to provide traffic managers and traffic participants with real-time, visualized traffic situation displays, traffic guidance strategies, and personalized travel services based on traffic situation prediction maps.

[0011] In the above technical solution, the base map production module uses remote sensing satellite image data and a trained deep learning model to extract the road network; it uses UAV aerial image data and an image edge detection algorithm model to extract the road network; and it uses vehicle BeiDou GNSS trajectory data to extract the road network structure through trajectory clustering and rasterization models.

[0012] In the above technical solution, the calculation process of the weights corresponding to the extraction results of each road network extraction model includes:

[0013] A certain number of real road network samples are selected to form a sample dataset;

[0014] The sample dataset is divided into K mutually exclusive subsets; and the sample dataset is further divided into a training sample set and a validation sample set.

[0015] Based on the mutually exclusive subset of samples corresponding to each model, the accuracy dataset containing user accuracy and producer accuracy is obtained for each model under K-fold cross-validation.

[0016] Calculate the coefficient of variation for each model's accuracy dataset;

[0017] The coefficients of variation of each model's accuracy dataset are summed and normalized to obtain the weights corresponding to each model.

[0018] In the above technical solution, the base map production module extracts highway intersections based on the highway network data generated by weighted fusion of each extraction result; radar, camera, and traffic sign elements are added to the base map according to their positional relationship with the highway intersections and the road network, thereby realizing the construction and visualization of the highway traffic spatiotemporal service base map.

[0019] In the above technical solution, the process of extracting highway intersections includes: extracting all road centerlines from the fused road network and obtaining the intersections of the centerlines as candidate intersections; using the vehicle-mounted Beidou GNSS trajectory data to perform trajectory clustering analysis on the candidate intersections, filtering out false intersections caused only by positioning errors or data drift; verifying the location of the remaining intersections by combining the spatial characteristics of satellite remote sensing data and UAV image data, and finally identifying the real set of highway intersection nodes.

[0020] In the above technical solution, the traffic situation analysis module obtains the traffic situation prediction value at each coordinate position on the highway traffic spatiotemporal service base map by weighting the predicted traffic flow, vehicle speed and congestion index.

[0021] The process of detecting abnormal traffic events in the above technical solution includes:

[0022] For each highway intersection, the system expands outward with different set radii to form different radiation zones corresponding to each highway intersection.

[0023] For each radiation area, kernel density analysis is performed at different time scales based on its corresponding traffic situation prediction value to obtain the time and location where the probability of anomalies in the radiation area is greater than a set threshold.

[0024] In the above technical solution, the decision support and personalized service module visualizes the traffic situation map and prediction results for traffic managers; the intelligent decision engine automatically generates traffic guidance and control strategies based on the real-time traffic information and traffic situation prediction results, combined with changes in vehicle speed, traffic flow fluctuations and accident alarms, and sends the strategies to road participants and relevant management terminals in the form of broadcast prompts, variable message sign information release, toll station control, traffic diversion, etc., to implement highway traffic operation control decisions.

[0025] In the above technical solution, the decision support and personalized service module targets traffic participants by sharing traffic management decision information to users' mobile devices through the platform's Internet information publishing middleware. Based on a travel preference model established by combining users' historical travel data, it provides users with personalized travel route planning suggestions and traffic information push services. At the same time, it receives real-time traffic information or personalized demand requests from users and uploads them back to the data warehouse to update the highway traffic spatiotemporal service base map and the prediction model in the traffic situation analysis module.

[0026] This invention provides a method for providing spatiotemporal services for highway traffic based on multi-source heterogeneous data fusion, comprising the following steps:

[0027] The data collected includes satellite remote sensing data, UAV imagery data, vehicle-mounted BeiDou GNSS trajectory data, and highway traffic data collected by roadside equipment.

[0028] By preprocessing, matching, integrating and representing the collected data, a hierarchical highway traffic spatiotemporal service data warehouse is constructed to achieve unified management of multi-source heterogeneous data.

[0029] Based on the data warehouse, deep learning, edge detection and trajectory analysis methods are used to extract road network information, and weighted fusion is performed based on the weights corresponding to each extraction result to generate a spatiotemporal service base map of highway traffic with highway intersections as key nodes.

[0030] Time series hierarchical analysis and convolutional neural network prediction are performed on traffic flow, vehicle speed and congestion index in the data warehouse. Based on the spatiotemporal service base map of highway traffic, cluster analysis is performed on the radiation area with highway intersection as the reference point to detect abnormal traffic events and form a traffic situation prediction map.

[0031] Based on traffic situation prediction maps, we provide traffic managers and traffic participants with real-time, visualized traffic situation displays, traffic guidance strategies, and personalized travel services.

[0032] The beneficial effects of this invention are as follows: It fully leverages the advantages of "air-space-ground" sensing technology in intelligent transportation, and combines it with data fusion management technologies and methods to create a hierarchical, secure, and efficient data warehouse. Based on multi-source heterogeneous highway traffic data and machine learning technology, it constructs a thematic highway base map for a highway traffic spatiotemporal service platform and system with intersections as key nodes, possessing advantages such as high accuracy and complete data structure. It also constructs a dynamic highway traffic temporal and spatial situation clustering analysis system, achieving comprehensive consideration of traffic conditions in both time and space dimensions, as well as information visualization within the base model of the highway traffic spatiotemporal service platform. This platform and method are fully geared towards traffic managers and individual traffic users, enabling information sharing and timely data acquisition between them.

[0033] Furthermore, the base map production module of this invention utilizes remote sensing satellite imagery (using a deep learning model), UAV imagery (using an edge detection algorithm), and vehicle BeiDou GNSS trajectory data (using trajectory clustering and rasterization methods) to extract road network information, thereby improving the accuracy of road network extraction and the integrity of the data structure. The multiple extraction methods complement each other, making full use of their respective advantages to generate high-precision road network data with a coherent topological structure, solving the problem of insufficient extraction accuracy in existing technologies. By using different data sources to verify and supplement each other, the road network information is ensured to be both comprehensive and accurate, thus providing a solid foundation for subsequent spatiotemporal situation analysis.

[0034] Furthermore, in the base map production module of the present invention, the weight calculation of the model results extracted from each road network is carried out as follows: real road network samples are selected and divided into K mutually exclusive subsets. K-fold cross-validation is used to obtain the user accuracy and producer accuracy of each model. After calculating the coefficient of variation and normalizing, the weight of each model is finally determined. Through an adaptive weighted fusion method, the prediction results of different models are automatically assigned appropriate weights according to their accuracy, thereby improving the accuracy and consistency of the overall road network fusion. K-fold cross-validation and statistical indicators (such as the coefficient of variation) are used to objectively assign weights to the models, making the base map production process fully open and feasible, and in line with the conventional technical requirements in this field.

[0035] Furthermore, the base map creation module of this invention extracts highway intersections from the weighted and fused road network data, and adds elements such as radar, cameras, and traffic signs to the base map according to their spatial locations, thereby realizing the construction and visualization of the base map. By extracting the key node of the intersection and overlaying traffic facility information, the resulting base map not only has a detailed road structure, but also intuitively reflects the important facilities in the road network, meeting the needs of traffic management and services. It realizes the organic combination of static road network and dynamic traffic element information, providing an intuitive and reliable spatial foundation for subsequent traffic situation analysis and decision support.

[0036] Furthermore, this invention extracts the intersections of all road centerlines from the fused road network as candidate intersections, then uses vehicle-mounted BeiDou GNSS trajectory data for trajectory clustering analysis to filter out false intersections, and combines this with spatial feature verification from satellite and UAV imagery to finally obtain a true set of intersection nodes. By verifying through multiple data sources and filtering out false intersections caused by positioning errors through cluster analysis, this ensures that the extracted intersections truly reflect the key nodes of the road network, further improving the accuracy of subsequent traffic situation prediction. The true intersection nodes serve as the benchmark for spatial analysis, helping to more accurately construct radiation areas during anomaly detection and forming a scientific spatial clustering analysis system.

[0037] Furthermore, the traffic situation analysis module of the present invention utilizes the predicted traffic flow, vehicle speed, and congestion index to obtain the traffic situation prediction value for each coordinate position on the highway traffic spatiotemporal service base map through weighted calculation; the weighted calculation for each coordinate position makes the numerical distribution of traffic situation more refined and more realistically reflects the local traffic state; the fine-grained traffic situation prediction value can provide managers with more accurate traffic condition distribution information, which helps to identify possible congestion or accident risk areas in advance.

[0038] Furthermore, this invention establishes different radii outward from each highway intersection to form different radiation zones. For each radiation zone, kernel density analysis is performed at different time scales to obtain the times and locations where the probability of anomalies is greater than a set threshold. By constructing radiation zones using intersections as reference points, the diffusion effect of traffic flow around intersections can be captured. Through multi-scale kernel density analysis, traffic anomalies can be detected in a timely manner, significantly improving the accuracy of accident or congestion warnings. High-risk areas and time periods can be identified in advance, providing a basis for traffic managers to implement preventive measures, thereby effectively reducing the probability of accidents and alleviating traffic congestion.

[0039] Furthermore, the decision support and personalized service module of this invention is designed for traffic managers. Through intuitive visualization and automated decision generation, it enables traffic managers to quickly grasp changes in road conditions, formulate emergency measures in a timely manner, and effectively respond to traffic emergencies. The multi-channel distribution of traffic guidance information ensures that information is quickly transmitted to various intersections and traffic participants, optimizes resource allocation, and improves the overall efficiency of traffic management.

[0040] Furthermore, the decision support and personalized service module of this invention targets traffic participants, enabling information sharing between traffic managers and users: users receive traffic decisions and road condition information in real time through mobile terminals, improving the timeliness and coverage of information transmission; combined with user travel history and real-time feedback, the system can continuously optimize route planning and traffic information push, realizing user-centric personalized travel services; user feedback data is used to update traffic situation analysis models and base map data, continuously improving the platform's prediction accuracy and response capabilities, thereby achieving intelligent dynamic traffic management. Attached Figure Description

[0041] Figure 1 This is a diagram of the overall architecture of the platform and system of this invention.

[0042] Figure 2 This is a schematic diagram illustrating the fusion and management of multi-source heterogeneous data in highway transportation.

[0043] Figure 3 This is a schematic diagram of the road network intersection extraction of the highway traffic infrastructure base.

[0044] Figure 4 This is a schematic diagram of the spatiotemporal situation data of highway traffic. Detailed Implementation

[0045] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments to facilitate a clear understanding of the present invention, but these descriptions do not constitute a limitation on the present invention.

[0046] Example 1

[0047] like Figure 1As shown, this invention provides a highway traffic spatiotemporal service platform based on multi-source heterogeneous data fusion, comprising:

[0048] The data acquisition module is used to collect data including satellite remote sensing data, UAV imagery data, vehicle-mounted BeiDou GNSS trajectory data, and highway traffic data collected by roadside equipment.

[0049] The data fusion management module is used to construct a hierarchical highway traffic spatiotemporal service data warehouse by preprocessing, matching, integrating and representing the collected data, so as to realize the unified management of multi-source heterogeneous data.

[0050] The base map creation module is used to extract road network information based on the data warehouse using deep learning, edge detection and trajectory analysis methods, and to perform weighted fusion based on the weights corresponding to each extraction result to generate a spatiotemporal service base map of highway traffic with highway intersections as key nodes.

[0051] The traffic situation analysis module is used to perform time-series hierarchical analysis and convolutional neural network prediction of traffic flow, vehicle speed and congestion index in the data warehouse. Based on the spatiotemporal service base map of highway traffic, it combines cluster analysis with the radiation area based on highway intersections to detect abnormal traffic events and form a traffic situation prediction map.

[0052] The decision support and personalized service module is used to provide traffic managers and traffic participants with real-time, visualized traffic situation displays, traffic guidance strategies, and personalized travel services based on traffic situation prediction maps.

[0053] Preferably, the multi-source data acquisition module generates raw highway traffic data from various sources, including high-resolution remote sensing satellite imagery, UAV aerial imagery, trajectory data generated by vehicle BeiDou GNSS terminals, and dynamic monitoring data such as vehicle speed, traffic flow counts, and video images collected by roadside cameras and radar. After preprocessing such as format conversion and noise filtering, these high-altitude, aerial, and ground-based data are output in a standardized manner. Subsequent use: The multi-source data output by the acquisition module is transferred to the data fusion management module as the input basis for fusion processing. For example, remote sensing imagery and UAV imagery will be used for road network extraction in subsequent base map production, and GNSS trajectory data and camera traffic flow data will be used for traffic flow feature calculation in situation analysis.

[0054] The data fusion management module generates a cleaned, aligned, and integrated comprehensive highway traffic database. This module unifies the format, aligns temporally and spatially, and eliminates redundancy in heterogeneous data from satellites, drones, vehicle terminals, and roadside equipment, outputting a structured fused dataset, which is stored in a cloud data warehouse (divided into different layers, such as raw data, processed data, and analytical data). Subsequent uses: The fused high-quality data serves as a unified data source for subsequent modules. On one hand, the base map creation module retrieves fused remote sensing imagery, aerial imagery, and trajectory data from the data warehouse for refined road network extraction and feature overlay. On the other hand, the situation analysis module extracts fused real-time and historical traffic indicator data (such as flow rate and speed) from the warehouse for calculating current traffic conditions and training predictive models. Therefore, the data fusion management module transforms previously scattered data into high-quality input that can be directly used for base map creation and situation analysis, improving the efficiency and accuracy of subsequent processing.

[0055] The base map generation module generates high-precision highway traffic spatiotemporal service base map data. This base map includes detailed highway network structure (road centerlines and connections represented in vector form), extracted key nodes of road intersections, and information on traffic sensing devices and traffic signs superimposed on it. The base map data contains both static road geometry and topology information, as well as information on the location of related traffic infrastructure. The base map serves as the foundation for situation analysis and decision-making visualization. In the situation analysis module, the base map provides a spatial reference framework—analyzed traffic flow, congestion, and other indicators are mapped onto the corresponding road segments or nodes on the base map, thus producing an intuitive traffic situation distribution map (such as a rasterized display of congestion levels). In particular, intersection nodes are used as the benchmark unit for spatial aggregation and anomaly detection in situation analysis, identifying high-incidence areas of anomalies by comparing traffic conditions in different intersection areas. Furthermore, in the decision support module, the base map is presented as a visualization background on the monitoring platform, assisting traffic managers in intuitively viewing real-time traffic conditions and prediction results, and deploying control measures more accurately based on the location of sensors on the base map. The high precision and completeness of the base map ensured that the subsequent situational analysis results accurately corresponded to the actual road environment.

[0056] The situation analysis module generates dynamic situation information and prediction results for highway traffic, including currently calculated traffic indicators (such as real-time traffic flow, average vehicle speed, and congestion index for each road segment), traffic condition prediction data for future periods, and detection and alarm information for abnormal events (accidents, abnormal congestion). In terms of format, this data may be represented as a set of numerical indicators, a traffic state matrix or raster map output by the prediction model, and a list of events marking abnormal locations. The results of the situation analysis will be directly used for decision support and user services. In the decision support module, the analyzed traffic situation data will be displayed to traffic managers in the form of charts and map overlays to help them understand the road network operation status in the next few hours or days; at the same time, the location and severity information of abnormal events provides a basis for managers to formulate emergency plans or allocate resources. In addition, the situation prediction results are also provided to ordinary traffic participants through the personalized service module, such as displaying predicted road conditions and alerting users to accidents ahead in navigation applications. In short, the situation analysis module serves as a bridge between the upper and lower levels: it uses basic data provided by the data warehouse and base map to perform calculations and feeds the resulting high-level information to the decision-making and service layers, driving the formulation of corresponding traffic control strategies and the adjustment of travel plans.

[0057] The decision support module generates data on decision information and control instructions for highway traffic management. Specifically, this includes a visualized traffic situation panel (a monitoring screen generated by integrating base maps and real-time / predictive traffic data), traffic control strategy schemes provided by the intelligent decision engine, and final decision commands or notifications (such as diversion instructions for congested road sections, warning broadcasts for accident-prone road sections, etc.). This output data includes both content displayed on the human-machine interface and control parameters for system execution or broadcasting. The output of the decision support module directly impacts traffic managers through the platform interface—allowing them to confirm or adjust control strategies; it is also transmitted to road users and related equipment via communication networks: for example, publishing control instructions on highway variable message signs or pushing them to navigation software to guide vehicles to detours. This decision information is also shared with the personalized service module, enabling ordinary users to promptly learn about the management department's control measures (such as the closure of a toll station or traffic control measures on a certain road section), thereby adjusting their travel plans. Furthermore, feedback during the decision-making process (such as data on whether traffic congestion has been alleviated after the implementation of control measures) is returned to the situation analysis module as new input, further enriching the data on the effectiveness of management interventions in the data warehouse, providing a reference for the next decision, and forming a closed-loop decision optimization process.

[0058] The personalized service module generates data in two types: customized service information for individual traffic participants and user feedback data. The former includes personalized travel route suggestions, real-time traffic alerts (such as accident warnings and detour suggestions for congested sections), and estimated arrival time updates, all generated based on the user's current location, destination location, and preferences. These are typically presented as messages or notifications. The latter refers to travel feedback data uploaded by user terminals, such as user-reported traffic conditions, route preferences, and destination change requests. Personalized service information is directly provided to end users to help them optimize their travel decisions and represents the final manifestation of the platform's external services. Once user feedback data is sent back to the platform, it enters the front end of the data flow: on one hand, the real-time traffic feedback supplements the multi-source data acquisition layer, enriching the perception of the current traffic status along with sensor data from cameras and other sources, improving the accuracy of situational analysis; on the other hand, long-term user feedback data (such as travel habits and route choices) can be stored in a data warehouse to update user preference models and improve the personalization of future route recommendations. This feedback mechanism enables the platform to continuously calibrate its analytical models and service content, further improve the real-time updates of the base map and situational analysis, provide more accurate information support for managers and other users, and form a data chain that continuously improves from data collection and analysis to decision-making and user application.

[0059] Specifically, the data acquisition module fully leverages the advantages of "space-air-ground" sensing technology in intelligent transportation. By acquiring high spatial resolution remote sensing image data, such as satellite imagery (e.g., WorldView, QuickBird, and domestic Gaofen series), and performing preprocessing including radiometric calibration, atmospheric correction, geometric correction, cloud (snow) removal, fusion, mosaicking, and cropping, as well as image enhancement operations such as histogram equalization and contrast stretching, a high-quality image set is obtained.

[0060] Considering the limitations of satellite imagery spatial resolution, it performs well in extracting main road networks, but its extraction quality for non-main road networks is poor. Therefore, UAV imagery information was further acquired. This data was obtained using a UAV equipped with a high-resolution camera, which conducted aerial photography according to a set flight path and overlap (typically 60-80% forward overlap and 20-40% lateral overlap). Further preprocessing operations such as image stitching, orthorectification, and image enhancement were then performed.

[0061] To obtain dynamic information on highway traffic, based on the acquired highway network information, vehicle-mounted BeiDou GNSS terminals are used to acquire vehicle trajectories. The hardware required for this process mainly includes: a vehicle-mounted BeiDou GNSS terminal for receiving BeiDou satellite signals and performing positioning calculations; an antenna for receiving satellite signals, usually integrated with or external to the terminal; a power supply for the terminal; a communication device for transmitting the data collected by the terminal to a server; and a storage device for locally storing the trajectory data.

[0062] For the large amount of vehicle trajectory data obtained using vehicle-mounted BeiDou GNSS terminals, the first step is to filter vehicle trajectories, i.e., filter vehicle trajectory data that are always in a state close to traffic flow; then, remove discrete trajectory points, i.e. remove trajectory points that do not have a set of similar trajectory points; finally, encrypt the trajectory points, i.e., use linear encryption to improve the level of trajectory aggregation in order to obtain overall vehicle trajectory information.

[0063] Specifically, the data fusion management module, the management module of the entire platform and system, utilizes satellites, drones, vehicle-mounted BeiDou GNSS terminals, cameras, and radar highway traffic equipment to acquire spatiotemporal data and information such as high spatial resolution images, trajectories, speeds, pictures, and videos. To address the issues of information redundancy, data heterogeneity, time asynchrony, and location misalignment among the acquired multi-source highway traffic data, further data fusion and management are implemented.

[0064] Data fusion technology typically includes data preprocessing, data matching, data integration, and data representation. Data preprocessing involves cleaning and transforming data formats to enable further analysis. Data matching involves identifying relationships between data items from different sources. Data integration combines the matched data to provide a unified view. Data representation ensures that the integrated data can be intuitively understood and manipulated by end users. Furthermore, the development of data fusion technology relies on machine learning and artificial intelligence algorithms, which can help automate complex tasks in the data fusion process, such as pattern recognition and predictive analytics. Utilizing these advanced technologies, highway traffic management platforms can provide more accurate and timely information, thereby supporting more effective management decisions and operations. The main framework is as follows: Figure 2 As shown.

[0065] Specifically, the base map creation module is the foundation for building a highway traffic spatiotemporal service platform and system. It mainly includes three parts: acquiring road network information, extracting key intersection nodes, and adding and visualizing typical highway elements. Its main workflow is illustrated below. Figure 3 As shown, the specific steps include:

[0066] (1) Obtaining road network information

[0067] To address the issue of obtaining road network information using satellite imagery data sources, a supervised learning approach is employed, combining existing road network sample sets and a deep learning model for highway network extraction, to achieve planar extraction of the road network. The main process and content are shown in Table 1.

[0068] Table 1. Satellite-based road network information extraction process and content

[0069]

[0070] Aerial photography of a specific area's road network using drones was employed to acquire and process the images. Then, edge detection algorithms were used to extract the road network. The road network extraction process and details are shown in Table 2.

[0071] Table 2. Process and Content of Drone Extraction of Road Network Information

[0072]

[0073]

[0074] In addition, the main process and content of obtaining road network data using trajectory information acquired by vehicle-mounted BeiDou GNSS terminals are as follows:

[0075] ① Calculate the driving direction angle of all trajectory points, then take into account the position and driving direction to obtain a set of similar trajectory points, and calculate the offset distance generated by each similar trajectory point;

[0076] ② The average of the summation along the offset direction is used to obtain the offset distance of the trajectory point to be offset. When the average offset distance of all trajectory points is greater than the threshold, the driving direction angle of the trajectory point is recalculated based on the offset coordinates until the average offset distance is less than the threshold.

[0077] ③ Based on the results of trajectory clustering, trajectory points that are not clustered are removed to obtain trajectory data that can reflect the road structure. Then, the road network is extracted using the raster digitization method.

[0078] The road network results extracted from remote sensing satellite imagery, UAV imagery, and BeiDou GNSS terminals differ to some extent. To integrate these different road network extraction methods, reduce discrepancies, and improve road network accuracy, an adaptive weighted fusion strategy is adopted to fuse the three road network extraction results. The calculation of the adaptive weighted fusion is as follows:

[0079] P = W1 × P s +W2×P t +W3×P r (1)

[0080] In the formula, P is the fused prediction value; P s P t and P r These are three road network extraction models (deep learning model, image edge detection algorithm model, trajectory clustering and rasterization model); W1, W2 and W3 are the weights of the three road network extraction models, and the sum of W1, W2 and W3 is 1.

[0081] To automatically determine the weight values ​​of the three road network extraction models, a certain number of real road network samples and the classification results of the corresponding road network extraction model locations are selected by combining high-resolution imagery and manual screening.

[0082] Generally, the confusion matrix method is used to obtain the user accuracy and producer accuracy of the classification model; the formulas for calculating user accuracy and producer accuracy are as follows:

[0083] UA=TP / (TP+FP) (2)

[0084] PA=TP / (TP+FN) (3)

[0085] Among them, UA and PA refer to User Precision (measured by the proportion of correctly predicted categories out of the total number of predictions for that category) and Producer Precision (measured by the proportion of samples correctly classified in the actual category); TP refers to the number of road network samples correctly classified, FP refers to the number of samples misclassified as road networks, and FN refers to the number of samples misclassified as other categories.

[0086] In this embodiment, to make the calculated accuracy metrics more representative and reduce the impact of sample imbalance, K-fold cross-validation is further employed to obtain the user accuracy and producer accuracy of the classification model. The main steps include the following:

[0087] ① Divide the sample data set D into K mutually exclusive subsets:

[0088] D = D1∪D2∪…∪D K (4)

[0089] ② Divide the sample set D, which is divided into K mutually exclusive sub-sample sets, into K classification model training sample sets and validation sample sets in order. The training sample set consists of K-1 sub-sample sets, and the validation sample set consists of 1 sub-sample set other than the training sample set.

[0090] For each classification model, the user accuracy (UA) and producer accuracy (PA) corresponding to the three road network extraction models are calculated according to formulas (2) and (3) for the training sample set and validation sample set respectively;

[0091] ③Calculate the user accuracy and producer accuracy of the three road network extraction models for each mutually exclusive subset of K-fold cross-validation results:

[0092]

[0093] The above provides a dataset combining user and producer accuracy for calculating three road network extraction models (deep learning model, image edge detection algorithm model, trajectory clustering and rasterization model), which is X. s :[X s1 X s2 , ..., X sm ], X r : [X r1 X r2 , ..., X rm ], X t :[X t1 X t2 , ..., X tm ], X ni It includes a road network extraction model n, and the user accuracy and producer accuracy for the i-th mutually exclusive subsample set.

[0094] Using this dataset and the information weighting method, the weights of three classification models are calculated. The information weighting method is an objective weighting method that uses the coefficient of variation of the data to assign weights. A larger coefficient of variation indicates more information carried, and therefore a larger weight. The calculation process is as follows:

[0095] ① Calculate the mean and standard deviation:

[0096]

[0097] Where m refers to the number of datasets combining user precision and producer precision, and n = s, r, and t represent the three types of datasets respectively. SD represents the average value. n Represents standard deviation.

[0098] ② Calculate the coefficient of variation (CV), which is the ratio of the standard deviation to the mean:

[0099]

[0100] ③ Calculate the weights, which are obtained by summing and normalizing the coefficients of variation:

[0101]

[0102] Where k = 3.

[0103] (2) Extraction of key nodes at intersections

[0104] Considering that the accident rate at road intersections is generally higher than at non-road intersections, it is necessary to extract the location and spatial extent of intersections to better serve highway traffic operations. Based on the above processing, high-quality road network data can be obtained. To extract typical road network nodes (intersections), further extraction of road network intersections is achieved by combining trajectory features. The specific details are as follows:

[0105] Based on the road network plan and edge information, road centerlines are extracted, and the intersections of these centerlines are extracted and marked as candidate road intersections. The specific process mainly includes:

[0106] ① Filter the tangent points P1 and P2 of two parallel tangent lines in the nearest neighborhood of the road network edge, and find the center point (Xi, Yi) of the line connecting the two tangent points;

[0107] ② Connect all center points (Xi, Yi) to generate the road centerline CL;

[0108] ③ Based on the center lines CL of all roads, obtain the intersection points CP of all roads.

[0109] Trajectory data clustering is used to remove false intersections and to filter intersection locations with high clustering of trajectory points near candidate road intersections. The specific process mainly includes:

[0110] ① It is necessary to filter vehicle trajectories, that is, to filter the trajectory data of vehicles that are always moving in a state close to traffic flow.

[0111] ②Remove discrete trajectory points, that is, remove trajectory points that do not have a set of similar trajectory points;

[0112] ③ Encrypt the trajectory points, that is, use linear encryption to improve the trajectory aggregation level;

[0113] ④ Considering that true intersections usually involve the convergence and dispersion of multiple trajectories, while pseudo-intersections are often caused by positioning errors, data drift, etc., we use cluster analysis to analyze the trajectory density near the intersection and combine it with spectral and texture features obtained from remote sensing satellite images and UAV images to identify and remove these trajectory intersections caused by accidental factors.

[0114] (3) Visualization of typical highway elements

[0115] Based on the extraction results of the above road network and its key nodes, with intersections as the base points of the highway traffic spatiotemporal service map, typical elements of highway traffic spatiotemporal services, such as radar, cameras, and traffic signs, are matched and added to the base map to realize the visualization of the base of the highway traffic spatiotemporal service platform, so as to facilitate the updating of real-time base map information and the efficient analysis and management of highway traffic spatiotemporal services.

[0116] By combining WebGL technology, a 3D visualization technology based on JavaScript and OpenGLES 2.0, and leveraging its rich API interfaces, the OpenGL interface is used for 3D rendering, and the graphics are drawn using the Canvas tag built into HTML5, thereby achieving smooth loading of the base model of the highway traffic spatiotemporal service platform.

[0117] Based on the fusion and management of heterogeneous highway traffic data, the situation analysis module conducts comprehensive monitoring and prediction of the temporal and spatial situation of highway traffic in order to achieve overall control over the regional highway traffic situation.

[0118] The situation analysis module analyzes the spatiotemporal situation of highway traffic using the following steps:

[0119] ① First, filter the main data needed in the data warehouse, collect data from various sensors, monitoring equipment, and vehicle Beidou GNSS terminals, and calculate key indicators including traffic flow, speed, and congestion index;

[0120] ② Regarding the temporal trends of highway traffic, the data is first divided into different time levels: daily, monthly, and quarterly. Combining time series analysis and convolutional neural network models, traffic flow, speed, and congestion are modeled to predict traffic conditions over a future period. The specific formulas included are as follows:

[0121] The input data is represented as follows:

[0122] X t =[X1(t),X2(t),…,X n (t)] (11)

[0123] Among them, X t The input feature vector at time t contains n features, namely the key indicators of traffic flow, speed, and congestion index in ①; X i (t) represents the observed value of the i-th feature at time t.

[0124] Convolution operations extract local time series feature representations:

[0125] Z t =X t *W+b (12)

[0126] Where W is the convolution kernel and b is the bias term.

[0127] Activation function representation:

[0128] A t =ReLU(Z) t (13)

[0129] Among them, At It is the output after activation; the ReLU function is the activation function.

[0130] A pooling layer (taking max pooling as an example) used to reduce data dimensionality while retaining key features is represented as follows:

[0131] P t =MaxPool(A t (14)

[0132] Among them, P t This is the output after pooling.

[0133] Fully connected layer representation:

[0134] Y t =W fc *P t +b fc (15)

[0135] Among them, Y t This refers to the prediction of traffic conditions at future moments, W fc b represents the weights of the fully connected layer. fc This is a bias term.

[0136] ③ Based on the predicted key indicators such as traffic flow, speed, and congestion index, a comprehensive traffic situation evaluation model is further constructed, and the predicted results are expressed on the road network in the form of different values ​​to construct a traffic situation prediction grid map:

[0137] S=W1*A+W2*(1-B)+W3*C (16)

[0138] Where S∈[0,1], the larger the value, the worse the traffic situation; A, B and C refer to the normalized traffic flow, speed and congestion index prediction data; W1, W2 and W3 refer to the weight values ​​corresponding to the traffic flow, speed and congestion index prediction data.

[0139] R = S i ,i∈(x i ,y i (17)

[0140] Where R refers to the final road network traffic situation prediction grid; S i This refers to the traffic situation prediction value at location i; x i ,y i This refers to the coordinates of position i.

[0141] ④ Regarding the spatial pattern of highway traffic, taking highway intersections as reference points, and expanding outwards with different radii, the corresponding radiation area for each highway intersection is determined. The formula can be:

[0142]

[0143] Where r refers to the distance of any coordinate position within the radiation area from the intersection; F refers to the radiation area centered on the intersection. If any coordinate position is more than 2km away from the intersection, it is considered outside the intersection's radiation area. The 0.5km, 1km, and 2km in Formula 18 are manually set thresholds that can be adjusted according to different application scenarios. Furthermore, the radiation area of ​​an intersection can be divided into several zones according to actual needs, not limited to the three shown in Formula 18, or... Figure 4 The two shown in the image.

[0144] like Figure 4 As shown, two non-overlapping radiation areas are determined for each intersection based on r1 and r2. Coordinates whose straight-line distance from a specified intersection exceeds r1+r2 are determined to be outside the radiation range of that intersection.

[0145] ⑤ Perform comprehensive cluster analysis on the traffic situation prediction values ​​S at different times (day, month, season) and different spatial ranges (different radiation areas) to locate areas with high kernel density, in order to obtain the time and location when abnormal traffic situation is likely to occur in each radiation area in the future. The specific formula is as follows:

[0146] Z = {K} d,F ,K m,F ,K p,F} (19)

[0147] Where Z refers to the comprehensive cluster analysis result; K d,F ,K m,F ,K p,F These refer to the kernel density analysis results of different intersection radiation regions on daily, monthly, and seasonal time scales, respectively, and the formula for calculating K is as follows:

[0148]

[0149] Where n is the number of traffic observation points (locations) used for analysis within the radiation area to be calculated; h is the bandwidth parameter; and e represents the traffic situation prediction value at a certain location. i This represents the traffic situation forecast for all other locations within the radiation area excluding this location; K(e) represents the density estimate for this location. It is important to note that the traffic situation forecasts are time-series data, including traffic situation predictions for different future times.

[0150] The decision support and personalized service module correlates the intensity of traffic anomalies in different radiation areas with the distance to intersections and adds it to the base model of the highway traffic spatiotemporal service platform. This enables decision support and judgment for highway traffic managers and personalized service customization for participating users. The decision support for highway traffic spatiotemporal services mainly includes monitoring and management consoles, intelligent decision engines, and decision deployment.

[0151] The monitoring and management console is presented in the form of a web page and a large screen. It is the control interface for highway traffic managers to design scheduling operations, conduct scheduling monitoring and resource monitoring, and includes two roles: administrator and operator. The administrator has all operation permissions of the platform, while the operator can only monitor the system operation status and has limited management functions when the operation is abnormal.

[0152] The intelligent decision engine is a foundational model for a highway traffic spatiotemporal service platform based on intersections as reference nodes. It uses comprehensive spatiotemporal situation analysis results within a specific time domain and region, employing cutting-edge cloud computing and artificial intelligence algorithms to comprehensively evaluate factors such as vehicle speed, traffic volume changes, and traffic accidents. Based on real-time road conditions, real-time traffic flow, predicted road conditions, and estimated travel time under different scenarios, it performs intelligent analysis and judgment to formulate or generate traffic guidance strategies for various traffic environments (including emergencies), and completes guidance and control functions such as broadcasting, guidance dissemination, toll station entry and exit management, and traffic diversion management.

[0153] Decision-making and deployment is an effective way to implement highway traffic operation decisions and deployments. It is mainly sent to main road participants and traffic management personnel in the form of broadcasts to achieve normal and efficient highway traffic operation.

[0154] For personalized service customization for highway traffic participants, on the one hand, based on the management of Internet information publishing middleware, it connects with third-party services or mobile applications. By sharing the support decisions of highway traffic managers to mobile devices in a timely manner, and building a cloud-based user travel preference model based on users' historical travel data, and combining the acquired highway traffic decision data, it provides personalized travel routes and traffic information push. On the other hand, it can also provide real-time feedback on highway traffic conditions to the management level, so as to realize the real-time updating of information on the base model of the highway traffic spatiotemporal service platform.

[0155] Example 2

[0156] This invention provides a method for providing spatiotemporal services for highway traffic based on multi-source heterogeneous data fusion, comprising the following steps:

[0157] The data collected includes satellite remote sensing data, UAV imagery data, vehicle-mounted BeiDou GNSS trajectory data, and highway traffic data collected by roadside equipment.

[0158] By preprocessing, matching, integrating and representing the collected data, a hierarchical highway traffic spatiotemporal service data warehouse is constructed to achieve unified management of multi-source heterogeneous data.

[0159] Based on the data warehouse, deep learning, edge detection and trajectory analysis methods are used to extract road network information, and weighted fusion is performed based on the weights corresponding to each extraction result to generate a spatiotemporal service base map of highway traffic with highway intersections as key nodes.

[0160] Time series hierarchical analysis and convolutional neural network prediction are performed on traffic flow, vehicle speed and congestion index in the data warehouse. Based on the spatiotemporal service base map of highway traffic, cluster analysis is performed on the radiation area with highway intersection as the reference point to detect abnormal traffic events and form a traffic situation prediction map.

[0161] Based on traffic situation prediction maps, we provide traffic managers and traffic participants with real-time, visualized traffic situation displays, traffic guidance strategies, and personalized travel services.

[0162] Example 3

[0163] The present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the spatiotemporal service method for highway traffic based on multi-source heterogeneous data fusion as described in the above technical solution.

[0164] Example 4

[0165] The present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the spatiotemporal service method for highway traffic based on multi-source heterogeneous data fusion described above.

[0166] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0167] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0168] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0169] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0170] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

[0171] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

Claims

1. A spatiotemporal service platform for highway traffic based on multi-source heterogeneous data fusion, characterized in that: include: The data acquisition module is used to collect data including satellite remote sensing data, UAV imagery data, vehicle-mounted BeiDou GNSS trajectory data, and highway traffic data collected by roadside equipment. The data fusion management module is used to construct a hierarchical highway traffic spatiotemporal service data warehouse by preprocessing, matching, integrating and representing the collected data, so as to realize the unified management of multi-source heterogeneous data. The base map production module is used to extract road network information based on the data warehouse using deep learning, edge detection and trajectory analysis methods, and perform weighted fusion based on the weights corresponding to each extraction result to generate a highway traffic spatiotemporal service base map with highway intersections as key nodes. The traffic situation analysis module performs time-series hierarchical analysis and convolutional neural network prediction on traffic flow, vehicle speed, and congestion index data in the data warehouse. Based on the spatiotemporal service base map of highway traffic, it combines cluster analysis with radiation areas based on highway intersections to detect abnormal traffic events, generating a traffic situation prediction map. The decision support and personalized service module provides real-time, visualized traffic situation displays, traffic guidance strategies, and personalized travel services to traffic managers and users based on the traffic situation prediction map. The base map creation module uses remote sensing satellite imagery data and a trained deep learning model to extract the road network; it also uses UAV aerial imagery data and an image edge detection algorithm model to extract the road network. Using vehicle BeiDou GNSS trajectory data, the road network structure is extracted through trajectory clustering and a raster model; The calculation process for the weights corresponding to the extraction results of each road network extraction model includes: selecting a certain number of real road network samples to form a sample dataset; dividing the sample dataset into K mutually exclusive subsets; further dividing the sample dataset into training and validation datasets; based on the mutually exclusive subsets corresponding to each model, obtaining the precision dataset containing user precision and producer precision under K-fold cross-validation; calculating the coefficient of variation of each model's precision dataset; summing and normalizing the coefficients of variation of each model's precision dataset to obtain the weights corresponding to each model; The base map creation module extracts highway intersections based on the weighted fusion of various extraction results to generate highway network data; radar, camera, and traffic sign elements are added to the base map according to their positional relationship with the highway intersections and road network, realizing the construction and visualization of the highway traffic spatiotemporal service base map; The process of extracting highway intersections includes: extracting all road centerlines from the fused road network and obtaining the intersections of the centerlines as candidate intersections; using the vehicle-mounted BeiDou GNSS trajectory data to perform trajectory clustering analysis on the candidate intersections, filtering out false intersections caused only by positioning errors or data drift; verifying the location of the remaining intersections by combining the spatial features of satellite remote sensing data and UAV imagery data, and finally identifying the real set of highway intersection nodes; The traffic situation analysis module uses the predicted traffic flow, vehicle speed, and congestion index to calculate the predicted traffic situation value at each coordinate location on the highway traffic spatiotemporal service base map through weighted calculation; The process of detecting abnormal traffic events includes: for each highway intersection, expanding outwards with different set radii to form different radiation areas corresponding to each highway intersection; for each radiation area, performing kernel density analysis at different time scales based on its corresponding traffic situation prediction value to obtain the time and location where the probability of anomalies occurring in the radiation area is greater than a set threshold.

2. The highway traffic spatiotemporal service platform based on multi-source heterogeneous data fusion according to claim 1, characterized in that: The decision support and personalized service module is designed for traffic managers, visualizing the traffic situation prediction map. The intelligent decision engine automatically generates traffic guidance and control strategies based on real-time traffic information and traffic situation prediction results, combined with changes in vehicle speed, traffic flow fluctuations, and accident alarms. These strategies are then sent to road participants and relevant management terminals in the form of broadcast prompts, variable message sign information releases, toll station control, and traffic diversion, enabling the implementation of highway traffic operation control decisions.

3. The highway traffic spatiotemporal service platform based on multi-source heterogeneous data fusion according to claim 1, characterized in that: The decision support and personalized service module targets traffic participants and shares traffic management decision-making information with users' mobile devices through the platform's internet information publishing middleware. Based on a travel preference model built by combining users' historical travel data, it provides users with personalized travel route planning suggestions and traffic information push services. At the same time, it receives real-time traffic information or personalized demand requests from users and uploads them back to the data warehouse to update the highway traffic spatiotemporal service base map and the prediction model in the traffic situation analysis module.

4. A method for providing spatiotemporal services for highway traffic based on multi-source heterogeneous data fusion, characterized in that: The system, implemented based on any one of claims 1-3, includes the following steps: collecting data including satellite remote sensing data, UAV imagery data, vehicle-mounted BeiDou GNSS trajectory data, and highway traffic data collected by roadside equipment; constructing a hierarchical highway traffic spatiotemporal service data warehouse by preprocessing, matching, integrating, and representing the collected data to achieve unified management of multi-source heterogeneous data; extracting road network information based on the data warehouse using deep learning, edge detection, and trajectory analysis methods, and weighting and fusing the extracted results according to their respective weights to generate a highway traffic spatiotemporal service base map with highway intersections as key nodes; performing time-series hierarchical analysis and convolutional neural network prediction on traffic flow, vehicle speed, and congestion index in the data warehouse; and performing cluster analysis based on the highway traffic spatiotemporal service base map, combined with the radiation area based on highway intersections as reference points, to detect abnormal traffic events and form a traffic situation prediction map; and providing real-time, visualized traffic situation display, traffic guidance strategies, and personalized travel services to traffic managers and traffic participants based on the traffic situation prediction map.

Images