A roadside data mining method and system

Abstract

This application discloses a roadside data mining method and system, solving the problem that integrated vehicle-road-cloud transportation systems cannot integrate data from vehicles and roadside infrastructure. A roadside data mining method includes the following steps: acquiring first information and second information; the first information includes roadside sensing information acquired through roadside sensing devices; the second information includes terminal status information acquired through information and communication networks; determining the spatiotemporal information of traffic participants in the first and / or second information; generating a traffic participant feature attribute set, the traffic participant feature attribute set including the number, type, and spatiotemporal information of traffic participants, and data representing traffic events determined based on the above information. This application integrates roadside and vehicle-mounted data to form a city-level big data platform, supporting the development and optimization of various smart city applications and improving the level of intelligent urban management.

Description

Technical Field

[0001] This application relates to the field of roadside sensing technology, and in particular to a roadside data mining method and system. Background Technology

[0002] Vehicle-road-cloud integration technology is transforming the transportation ecosystem. With the deep integration of the Internet of Things, big data, cloud computing, and artificial intelligence, every network-connected vehicle becomes a mobile data generator, continuously contributing massive amounts of data, including location information, operational status, environmental perception, and even driver and passenger behavior patterns. The sheer scale and breadth of this data undoubtedly opens a new door to smart city governance, while also bringing unprecedented challenges and opportunities. How to efficiently utilize and properly store the data generated by these massive numbers of traffic participants is crucial for promoting the intelligent upgrading of smart city governance systems.

[0003] Existing technologies in vehicle-road-cloud integrated transportation cannot integrate data from vehicles and roadside infrastructure. Therefore, a solution is needed that can efficiently acquire, transmit, process, and analyze traffic data from multiple sources, including vehicles, roadside infrastructure, and mobile devices, to maximize the utilization of such data and promote intelligent governance and optimization of transportation systems in smart cities. Summary of the Invention

[0004] This application provides a roadside data mining method and system, which solves the problem that existing technologies in vehicle-road-cloud integrated transportation cannot integrate data from vehicles and roadside infrastructure.

[0005] In a first aspect, embodiments of this application provide a roadside data mining method, including the following steps:

[0006] Acquire first information and second information; the first information includes roadside sensing information acquired through roadside sensing devices; the second information includes terminal status information acquired through information and communication networks.

[0007] Determine the spatiotemporal information of traffic participants in the first and / or second information;

[0008] A set of traffic participant feature attributes is generated, which includes the number, type, and spatiotemporal information of traffic participants, as well as data representing traffic events determined based on the above information.

[0009] In one embodiment, the step of:

[0010] Acquire high-precision map data;

[0011] Data processing is performed based on the characteristic attribute set of traffic participants and high-precision map data to generate twin dynamic display, traffic flow prediction or accident early warning data.

[0012] In one embodiment, the information and communication network includes at least one of the following: V2X, personal mobile communication, and data communication network;

[0013] The terminal includes at least one of the following: vehicle-mounted terminal, traffic control terminal, and personal mobile terminal;

[0014] The vehicle terminal status information includes at least one of the following: information indicating the vehicle's position, speed, direction, or operating status;

[0015] Traffic control terminal status information includes at least one of the following: information indicating the position, speed, or direction of traffic participants;

[0016] Personal mobile terminal status information includes at least one of the following: information indicating the location, speed, or direction of a pedestrian or driver.

[0017] In one embodiment, data processing of the traffic participant attribute set includes at least one of the following:

[0018] Data cleaning, feature extraction, pattern recognition, outlier handling, data augmentation, or processing model optimization.

[0019] In one embodiment, the storage method is determined to be at least one of the following, depending on the data size or access requirements: cloud storage, edge storage, solid-state drive storage, distributed file system, or data storage.

[0020] In one embodiment, the first information, the second information, or the set of traffic participant characteristic attributes are transmitted back to the cloud, and the transmission method includes: transmission via network transmission tool, transmission via cloud storage, transmission via hard disk, or transmission via frame extraction and compression.

[0021] Secondly, embodiments of this application also provide a roadside data mining system for implementing the roadside data mining method described in any embodiment of the first aspect, comprising: a sensing device, an information transmission device, and a data processing device. The sensing device is used to acquire first information and second information. The information transmission device is used to transmit the acquired first and second information to the data processing device. The data processing device is used to determine the spatiotemporal information of traffic participants in the first and / or second information; and is also used to generate a set of feature attributes of traffic participants.

[0022] In one embodiment, the data processing device includes a first data processing device as an edge computing unit and / or a second data processing device as a server.

[0023] Furthermore, the data processing device includes a server and an edge computing unit;

[0024] The edge computing unit, located at a roadside node, is used to determine the spatiotemporal information of traffic participants in the first and / or second information; it is also used to generate a set of feature attributes for traffic participants. The server, located in the cloud, is used to acquire high-precision map data; it is also used to process data based on the set of feature attributes for traffic participants and the high-precision map data to generate twin dynamic displays, traffic flow predictions, or accident early warning data.

[0025] Thirdly, embodiments of this application also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect above.

[0026] Fourthly, embodiments of this application also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in the first aspect above.

[0027] The above-described technical solutions adopted in the embodiments of this application can achieve the following beneficial effects:

[0028] This application integrates roadside and vehicle-mounted data to form a city-level big data platform, supporting the development and optimization of various smart city applications and improving the level of intelligent urban management. It provides data support for traffic policy formulation and urban planning, and optimizes traffic management: by deeply mining roadside data, it accurately grasps traffic flow characteristics, including but not limited to traffic volume distribution, vehicle type ratio, and traffic speed fluctuations. With the help of video surveillance and AI analysis, it monitors road safety hazards in real time, providing scientific basis for urban traffic managers to more accurately control traffic light timing, optimize road design, and adjust public transportation routes, thereby alleviating traffic congestion and improving the overall road network efficiency. This application can also be used for emergency response and disaster management: in the event of emergencies such as traffic accidents or severe weather, it rapidly integrates roadside data to provide precise navigation for emergency rescue, accelerate response speed, and provide scientific guidance for post-disaster traffic recovery, ensuring public travel safety. Attached Figure Description

[0029] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0030] Figure 1 This is a system device diagram illustrating the application scenario of this application;

[0031] Figure 2 This is a flowchart illustrating an embodiment of a roadside data mining method provided in this application.

[0032] Figure 3Flowchart of a first information processing embodiment for a roadside sensing device using a camera;

[0033] Figure 4 A flowchart of a first information processing embodiment for a roadside sensing device using lidar;

[0034] Figure 5 This is a flowchart illustrating an embodiment of the roadside data mining method of this application, which includes the generation of traffic management data.

[0035] Figure 6 A schematic diagram of a system embodiment for roadside data mining provided in this application;

[0036] Figure 7 This is a structural diagram of the computer device in the embodiments of this application. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0038] Roadside data mining is a key method for optimizing transportation networks, improving urban planning quality, and enhancing the effectiveness of intelligent transportation systems. The solution described in this application covers strategies for data acquisition, processing, analysis, and addressing challenges to maximize data value and effectively handle difficult samples, using data mining and analysis to feed back into the creation of a smart transportation platform.

[0039] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0040] Figure 1 This is a system device diagram illustrating the application scenario of this application, including an on-board terminal 110, a roadside sensor 120, and a data processing device 130. The data processing device includes a server 130-1 and / or a roadside computing unit 130-2. The on-board terminal is a processing device or a collection of processing devices installed in a vehicle. For example, at least a portion of the on-board terminal may be an on-board computing unit, etc. This application does not specifically limit the type and number of on-board terminals.

[0041] Among them, the roadside sensor is a roadside sensing device / roadside multi-source fusion sensing system set at a predetermined location at the entrance of the traffic scene and inside the traffic scene. For example, it can be a device pile evenly distributed in the traffic scene, or it can be at least one of millimeter-wave radar sensor, lidar sensor and camera, etc. The embodiments of this application do not specifically limit the type of roadside sensor.

[0042] The data processing device can be a server or a roadside computing unit. For example, the data processing device can be a central server set up on a cloud platform, or it can be an edge computing unit deployed at key traffic nodes (such as intersections) in a traffic scenario.

[0043] Furthermore, the system also includes a high-precision map database 140. The system can perform visualization processing on the high-precision map, and can also visualize the current road and traffic participant perception information generated by the digital twin of the computer device, forming a visual interface. The visualization processing module can be integrated into the vehicle terminal or set up separately in other locations.

[0044] Furthermore, it also includes a display terminal 160; the display terminal is used to realize traffic management visualization.

[0045] The system may also include a data storage device (not shown in the figure) for storing data that the computer equipment needs to process. The data storage system can be integrated into the vehicle terminal or placed on a server or other network.

[0046] The system described in this application embodiment may also include the following data interfaces: a first information interface ① between roadside sensors and data processing equipment, a second information interface ② between vehicle-mounted terminals and data processing equipment, a traffic participant feature dataset interface ③ between the first data processing equipment (e.g., a roadside edge computing unit) and the second data processing equipment (e.g., a cloud server), a data transmission interface for a high-precision map database ④, and a traffic management data interface output by the data processing equipment ⑤.

[0047] In one embodiment of this application, the first data processing device outputs a traffic participant feature dataset, which is transmitted to the second data processing device via a communication network. The second data processing device receives the traffic participant feature dataset and high-precision map data, and outputs data processing results for traffic management.

[0048] In one embodiment of this application, the second information interface is bidirectional, that is, the data processing device collects terminal status information through the second information interface and then feeds back traffic management information through the second information interface to change the terminal status.

[0049] This application utilizes data processing equipment to process data from roadside sensors and vehicle-mounted terminals to generate high-value data. In the field of traffic vehicle-road-cloud perception, high-value data mainly refers to data types that can significantly improve traffic efficiency and safety, and promote the intelligent upgrading of smart transportation systems. Specific data is described in the examples below.

[0050] Figure 2 The flowchart of the roadside data mining method provided in this application includes steps 210 to 230.

[0051] Step 210: Obtain the first information and the second information.

[0052] The first information includes roadside sensing information acquired by roadside sensing devices. This first information includes the type, location, speed, and number of traffic participants sensed in the road application scenario.

[0053] The perceived information is acquired through roadside sensing devices installed on the road. The construction of roadside sensor equipment involves deploying high-definition cameras, radar, infrared sensors, magnetic induction coils, etc., at key intersections and road sections to monitor traffic flow, vehicle type, speed, and vehicle spacing. By deploying sensing devices on the roadside, information about traffic participants is collected for target detection, trajectory tracking, and abnormal event analysis. This establishes a smart twin management platform, providing decision-making information for managers, or transmitting detection results to vehicles to provide drivers with information about the surrounding driving environment and assist driving.

[0054] In one embodiment, when the traffic participant is a vehicle, the first information specifically includes traffic flow, vehicle type, driving speed, and vehicle spacing.

[0055] The second information includes terminal status information obtained through information and communication networks (ICT). This second information includes the terminal ID, the terminal-indicated location, speed, and direction. When a traffic participant carries a terminal, the terminal status information serves as the traffic participant's ID, along with real-time self-measured location, speed, and direction, and the planned location, speed, and direction, forming a planned route. When a traffic control facility is equipped with a control terminal, the terminal status information can serve as the traffic participant's location, speed, and direction indicated by the traffic control terminal.

[0056] In one embodiment, when the traffic participant is a vehicle, the second information specifically includes the vehicle's location, speed, direction, or operating status.

[0057] For example, the information and communication network equipment includes in-vehicle terminal equipment, which integrates a dashcam or OBD-II (On-Board Diagnostics) device with sensors such as GPS, accelerometer, gyroscope, and camera. These devices can record vehicle location, speed, direction, and other information in real time. In one embodiment, this includes vehicle operating status, driving behavior, or vehicle-to-vehicle communication. Vehicle operating status refers to monitoring information about the vehicle's condition during operation, such as fuel consumption, fault warnings, or remaining fuel. Driving behavior refers to the recording of commands issued to the vehicle during driving, such as steering, acceleration, braking, and the frequency of rapid acceleration or braking. Vehicle-to-vehicle (V2V) communication data refers to real-time communication data between vehicles, used to optimize vehicle performance, improve driving safety, support insurance actuarial calculations, and provide personalized services.

[0058] For example, the information and communication network includes a mobile device application service network. This involves developing or accessing personal mobile terminal applications that utilize GPS positioning to collect anonymous user travel data. With user consent, information such as travel routes, times, and speeds can be obtained.

[0059] For example, the information and communication network includes Wi-Fi and Bluetooth probes. Wi-Fi probes and Bluetooth beacons are deployed in high-traffic areas to indirectly estimate the flow of pedestrians, drivers, and vehicles by detecting changes in the MAC addresses of passing mobile terminal devices.

[0060] The second information can be sent from the terminal to the data processing device via information (IT) services. For example, the vehicle terminal device or personal mobile terminal device can upload the second information to an independent server or cloud device that provides navigation or traffic information guidance services via an application (APP), and the data processing device can download the second information from the independent server or cloud device.

[0061] The second information can also be sent from the terminal to the data processing device via a communication network (CT). For example, the vehicle-mounted terminal device or the personal mobile terminal device can directly send the second information to the data processing device. For example, the vehicle-mounted terminal can send the second information to the data processing device via V2X communication technology.

[0062] Step 220: Process the first information and / or the second information to determine the spatiotemporal information of traffic participants in the first information and / or the second information.

[0063] The spatiotemporal information of traffic participants refers to the location information of traffic participants during their movement and the time information corresponding to the location information.

[0064] Based on spatiotemporal information, the same target traffic participants in the first and second information are identified. Traffic participants are identified to determine their type and identity information. In the first information, the type, location, speed, and number of target traffic participants, along with the corresponding time information, are determined. In the second information, the terminal ID, terminal location, speed, and direction of the target participants, along with the corresponding time information, are determined. The above information is comprehensively analyzed, and based on the time information, the same traffic participants in the first and second information are identified. The first and second information for these same traffic participants are then merged.

[0065] In one embodiment, the data processing device performs data processing on the first information and / or the second information, including at least one of the following: data cleaning, feature extraction, pattern recognition, outlier handling, data augmentation, and processing model optimization.

[0066] The data cleaning process removes invalid data records. For example, it removes invalid, duplicate, and erroneous data records to ensure data quality.

[0067] The feature extraction process involves extracting specific features from the first and second information. For example, meaningful features such as peak traffic flow, average vehicle speed, and congestion duration can be extracted from the raw data.

[0068] The pattern recognition identifies traffic flow patterns. For example, machine learning algorithms are used to identify traffic flow patterns, predict traffic congestion trends, and analyze areas with frequent traffic accidents.

[0069] Outlier handling involves identifying and processing extreme or anomalous data. For example, statistical analysis can be used to identify and process extreme or anomalous data to prevent negative impacts on model training.

[0070] For certain special events (such as sudden accidents), the data augmentation process can employ data augmentation techniques or transfer learning methods to enhance the generalization ability of the model by utilizing data from similar scenarios.

[0071] For example, for certain special events (such as sudden accidents), data augmentation techniques or transfer learning methods can be used to enhance the generalization ability of the model by utilizing data from similar scenarios.

[0072] The processing model optimization optimizes the traffic model parameters based on the newly collected data to ensure that the model adapts to changes in traffic conditions.

[0073] Step 230: Generate a set of feature attributes for traffic participants.

[0074] The aforementioned set of traffic participant characteristic attributes includes the number, type, and spatiotemporal information of traffic participants, as well as data representing traffic events determined based on the above information. The number of traffic participants is the total number of all traffic participants within a designated area at a certain point in time or during a certain time period.

[0075] The types of traffic participants include vehicle types, pedestrians or drivers, non-motorized vehicles, etc.

[0076] The spatiotemporal information of the traffic participants includes the location of the traffic participants and the time corresponding to the location.

[0077] Furthermore, when identifying traffic participants, possible solutions include: acquiring captured information and perceived information of traffic participants within a designated area on the road; generating identification information of traffic participants based on the captured information; the captured information can be, but is not limited to, images of traffic participants captured by the camera. For example, vehicle identification information can be, but is not limited to, license plate information.

[0078] Based on the perception of traffic participant identifiers, the location of traffic participants corresponding to a given time can be determined. This further allows for the acquisition of spatiotemporal information about traffic participants.

[0079] Depending on the specific application, operations such as clustering, segmentation, and feature extraction can be performed on the spatiotemporal information of traffic participants in the first and / or second information.

[0080] Data representing traffic events includes, for example, whether an event has occurred at the current time and the type of event that has occurred. The traffic events include, for example, speeding, slow driving, illegal parking, pedestrians or drivers, driving against traffic, littering, lane changing, etc. Data representing traffic events is generated after spatiotemporal information analysis of the target traffic participants.

[0081] By analyzing the output video from the cameras and the output point cloud from the radar, traffic participant information is obtained. The analyzed information is represented as a set of traffic participant feature attributes, as shown in Table 1. The set of traffic participant feature attributes includes: traffic participant fusion attributes, traffic participant point cloud attributes, and traffic participant video attributes.

[0082] Table 1. Characteristic Attribute Set of Traffic Participants

[0083]

[0084]

[0085]

[0086]

[0087] It should be noted that while the content attribute types of traffic participant fusion, traffic participant point clouds, and traffic participant videos in the table above are the same, the attribute values ​​differ. For example, in the point cloud dataset, the color attribute can be 0; in the video dataset, information regarding location, velocity, scale, and coordinates (longitude to Z-axis coordinate attributes listed in the table) may be invalid or have low precision. Therefore, further data parsing is needed to extract valid or high-precision data from each dataset.

[0088] It should be further noted that in step 210 of the above embodiment, the data sources for the roadside perception scheme are mainly twofold: the first is the output video from the camera, and the second is the output point cloud from the radar. The road scene can be a road scene jointly acquired by camera photography and radar detection. The road scene includes static and dynamic scenes. Static scenes include, but are not limited to, roads, road equipment, and ETC entrances / exits, while dynamic scenes include, but are not limited to, vehicles and the status of road equipment. The following describes how to anonymize these two types of data, complete the parsing, and feed it into the algorithm model.

[0089] As one embodiment, the roadside sensing device is a camera that receives and parses the first information, such as... Figure 3 As shown, the specific steps include:

[0090] Steps 210-11: Obtain the video stream and decode the video.

[0091] Video detection algorithms are based on processing single images, so their input data format is usually image files, such as JPEG and PNG. Roadside perception algorithms need to detect objects in video in real time. The common approach is to decompose the video into a series of frames and then apply the YOLOv algorithm frame by frame for object detection.

[0092] For video stream acquisition, if it's from a webcam or other network source, the video stream can be retrieved using protocols such as RTSP (Real-Time Streaming Protocol) or HTTP. If it's a local video file, it can be read directly.

[0093] Video decoding, for example, using video decoding libraries (such as FFmpeg) to decode video streams and convert them into frames of images.

[0094] Steps 210-12: Perform target detection on each frame of the image.

[0095] For example, the image is processed, specifically preprocessed for each frame, such as scaling and normalization, to meet the requirements of the YOLOv model. The processed image is then input into the YOLOv model for object detection.

[0096] Step 210-13: Visualize the target detection results.

[0097] Acquire the target detection results for each frame of the image and perform visualization processing as needed (such as drawing bounding boxes, labels, etc.). You can choose to recombine the processed images into a video stream output or display them directly on the screen.

[0098] In one embodiment, in step 210, the roadside sensing device is a lidar, which receives and parses the first information, such as... Figure 4 As shown, the specific steps include:

[0099] Steps 210-21: Acquire radar point cloud data.

[0100] Common radar point cloud data formats include, but are not limited to:

[0101] ASCII text format: A simple text file where each line records information from a single dot, and fields are usually separated by commas or spaces.

[0102] Binary files: To improve efficiency, some radar data is stored in binary format, which is compact and suitable for fast processing.

[0103] Specific packet formats: Some radar systems may have their own proprietary formats, such as the PointCloud2 message type in ROS (Robot Operating System), which can carry additional metadata.

[0104] Standard file formats, such as .pcd (Point Cloud Data) files, are widely used point cloud file formats defined by PCL (Point Cloud Library).

[0105] Analyzing radar point cloud data:

[0106] Steps 210-22: Perform target detection on the point cloud data.

[0107] To load data, for ASCII text format, you can use Python's pandas library to read CSV files. For binary files, you need to read and parse the data according to the specifications provided by the radar manufacturer.

[0108] For example, .pcd files can be loaded using the PCL library or tools such as Open3D.

[0109] Furthermore, the data can be preprocessed, for example, by transforming the coordinate system of the data. Other examples include filtering out invalid points, such as points at infinity. Still others include data denoising.

[0110] Steps 210-23: Visualize the target detection results.

[0111] For example, use libraries such as mayavi and matplotlib to visualize point cloud data.

[0112] Figure 5 This is a flowchart illustrating an embodiment of the roadside data mining method of this application, which includes the generation of traffic management data. In this embodiment, based on steps 210-230, steps 240-250 are further included:

[0113] Step 240: Obtain high-precision map data;

[0114] High-precision maps, such as high-precision electronic navigation maps for roads, contain the geometry, attributes, and relationships of road networks, lane networks, road markings, and road facilities. They support the access of dynamic road data, assist in the autonomous and intelligent movement of road vehicles and the refined management of road traffic, and are electronic maps or datasets that can be used in conjunction with general navigation electronic maps.

[0115] The geographic information data in high-definition maps can be three-dimensional geographic information data, including geographic information data of any point on roads and road equipment. Alternatively, it can be two-dimensional data obtained by preprocessing the three-dimensional geographic information data using high-definition maps.

[0116] In one embodiment of this application, a high-precision map of the target area is stored in a high-precision map database 140. It is connected to a server and / or a roadside edge computing unit via a data transmission interface ④. When an operator retrieves the high-precision map data of the target area via a computer, a high-precision map image is generated on a computer terminal or a GUI connected to it.

[0117] In this embodiment, graphic information and attribute information such as latitude and longitude of roads and road equipment are obtained from high-precision map data. The status of traffic participants in the road scene can be displayed in real time at the corresponding location on the high-precision map, which facilitates the location tracking and management of traffic participants.

[0118] In one embodiment, traffic data, including geographic location and spatiotemporal big data, can also be obtained through the interface of the traffic management system. By combining GPS, Beidou and other satellite positioning data, accurate location services can be provided, and accurate location information and lane attributes can be obtained by collecting high-precision maps of the actual site.

[0119] Step 250: Process the data based on the traffic participant characteristic attribute set and high-precision map data to generate twin dynamic display, traffic flow prediction or accident early warning data in the high-precision map.

[0120] Regarding the dynamic display data of twins: A twin image of a traffic scene is generated, in which fused information of traffic participants is generated. This includes twin objects of traffic participants. A twin object refers to a visual representation of an object in the real physical world, including but not limited to static objects such as buildings, terrain features, and traffic facilities, as well as dynamic objects such as traffic participants (vehicles, pedestrians, or drivers), weather conditions, and traffic time. In this embodiment, the road scene is jointly acquired by camera capture and radar detection, and the vehicles, roads, and road equipment included in the traffic scene are twinned into the twin image, binding the identity information of traffic participants and their positions in the high-precision map. In this application, the twin image uses a high-precision map, and the road scene and the twin image correspond. The twin image of the high-precision map can display complex road conditions and the status of road equipment, facilitating management by road management personnel.

[0121] Regarding traffic flow prediction data, traffic flow at any location and time on a road is simulated and calculated using the characteristic attribute set of traffic participants in the road system and road information from high-precision maps. For example, by determining the characteristic attribute set of traffic participants perceived within a certain road range around an intersection and the road information from a high-precision map, the traffic flow change curve at that intersection over a set time period is simulated and calculated, serving as traffic flow prediction data.

[0122] Regarding accident early warning data, traffic accidents that will occur to traffic participants are predicted by analyzing relevant traffic participant information and traffic participant characteristic attribute sets. For example, the probability of a traffic accident occurring at any location and time on a road is calculated by simulating the traffic participant characteristic attribute set of the road system and road information in high-precision maps, and an alert is issued for locations and times with probabilities greater than a set threshold.

[0123] Figure 6 This is a schematic diagram of a system embodiment for roadside data mining provided in this application. It includes: a sensing device, a terminal device, an information transmission device, and a data processing device. The sensing device is used to generate first information, and the terminal device is used to generate second information. The information transmission device is used to transmit the acquired first and second information to the data processing device. The data processing device is used to determine the spatiotemporal information of traffic participants in the first and / or second information; it is also used to generate a set of feature attributes of traffic participants. The data processing device includes a first data processing device as an edge computing unit and / or a second data processing device as a server. At the application layer, the terminal device and the server device support traffic information management services.

[0124] Specifically, the sensing devices include a video AI camera 310, a LiDAR 320, and a surveillance camera 330; the data processing devices include an edge computing unit 340, a monitoring chassis 350, and a cloud platform (server) 360. Application layer devices include a terminal 370 and a cloud platform. Information transmission devices include wireless communication V2X systems, fiber optic transmission systems, and communication network equipment.

[0125] like Figure 6 As shown, by utilizing sensing devices such as cameras, millimeter-wave radar, and lidar, combined with roadside edge computing, the ultimate goal is to achieve instantaneous intelligent perception of traffic participants and road conditions on this road segment. Through V2X vehicle-to-infrastructure (V2X) technology, and following agreed-upon communication protocols and data interaction standards, information exchange and command control between people, vehicles, roads, and the cloud are realized. The relevant information interfaces include: a first information data interface ① between sensing devices and data processing devices; a second information data interface ② between terminal devices and data processing devices; and data interfaces ③ between different data processing devices. At least one server has a high-precision map interface ④, and at least one server has an output traffic management data interface ⑤.

[0126] The terminal 370 is used to acquire second information. The terminal includes at least one of the following: a vehicle-mounted terminal (i.e., a vehicle-side terminal), a traffic control terminal, and a personal mobile terminal. The status information of the vehicle-mounted terminal refers to the terminal installed on the vehicle, used to provide information about the vehicle's status, including: information indicating the vehicle's position, speed, direction, and operating status. The status information of the traffic control terminal includes at least one of the following: information indicating the position, speed, or direction of traffic participants, such as traffic lights, direction indicators, speed limit indicators, and other prompts. The status information of the personal mobile terminal includes at least one of the following: information indicating the position, speed, and direction of pedestrians or drivers.

[0127] The edge computing unit and / or cloud platform are used to determine the spatiotemporal information of traffic participants in the first information and / or the second information; the edge computing unit and / or cloud platform are also used to generate a set of feature attributes of traffic participants.

[0128] For example, both the edge computing unit and the cloud platform can be used to determine the spatiotemporal information of traffic participants in the first and second information, and to generate a set of feature attributes for traffic participants. Data processing is jointly performed by the edge computing unit and the cloud platform; the edge computing platform performs preprocessing, and then the cloud platform performs further processing.

[0129] For example, for computationally intensive data preprocessing, initial screening and processing can be performed in edge computing units, reducing the pressure on central servers and improving efficiency. Deploying intelligent edge computing units at key traffic nodes (such as intersections) enables preprocessing and analysis directly at the source of data generation, reducing the time and bandwidth requirements for data transmission back to the cloud. As another example, data collected by roadside sensing devices is transmitted to the edge computing unit via a first information interface for initial screening before being transmitted to the server. Encryption measures are implemented during data transmission to protect personal privacy and data security.

[0130] When both the edge computing unit and the cloud platform are used to determine the spatiotemporal information of traffic participants in the first information and generate a set of feature attributes for traffic participants, the data collected by the sensing device can be transmitted to either the edge computing unit or the server. For example, the roadside sensing device sends the first information to the edge computing unit or the cloud platform through the first information interface. For example, such as... Figure 3 As shown, the video AI camera and LiDAR send corresponding roadside perception information to the edge computing unit. The monitoring camera directly uploads the roadside perception information to the cloud platform (server) through the monitoring chassis.

[0131] When both the edge computing unit and the cloud platform are used to determine the spatiotemporal information of traffic participants in the second information and generate a set of feature attributes for traffic participants, the data generated by the terminal device can be transmitted to either the edge computing unit or the server. For example, the vehicle terminal sends the second information to the edge computing unit or the cloud platform through the second information interface.

[0132] Data from the vehicle-mounted terminal enters the server in two ways: 1) An interface is established between the vehicle-mounted terminal's app backend server and the roadside system's backend server. For example, the vehicle-mounted terminal uploads the second information to the backend server of a navigation or traffic information app via the app, and the cloud platform downloads the second information from the app's backend server. 2) Data is directly collected and entered into the roadside system's backend server via roadside communication equipment connected to the vehicle-mounted terminal. For example, the vehicle-mounted terminal sends the second information to an edge computing unit via V2X, and the edge computing unit uploads the second information to the cloud platform.

[0133] Real-time data transmission between the server and the vehicle terminal is primarily achieved through optimized network configuration, adoption of fast-response communication protocols, and reduction of data processing and transmission latency. Furthermore, for real-time data transmission, the protocol stack includes a priority control mechanism to ensure that critical data packets are transmitted first, thereby reducing latency and ensuring timely data delivery.

[0134] Data transmission between the edge computing unit and the vehicle-mounted terminal can be achieved in real time via a V2X wireless communication system. In the system of this application, it is understood that the edge computing unit can transmit the acquired second information data back to the server. The server can also distribute the acquired second information data to the edge computing unit. Data exchange between the server and the edge computing unit can be real-time data forwarding or can be achieved through a storage device.

[0135] The edge computing unit and the cloud platform (server) can work separately. For example, the division of labor can be determined based on the amount of data. When the workload of data collection and parsing exceeds a set threshold, the edge computing unit performs data processing, specifically to generate a set of traffic participant feature attributes. The cloud platform (server) then processes the traffic participant feature attribute set, specifically to acquire high-precision map data and perform data processing based on the traffic participant feature attribute set and the high-precision map data to generate twin dynamic display, traffic flow prediction, or accident early warning data.

[0136] For example, the edge computing unit receives the first information, determines the spatiotemporal information of the traffic participants through the first information, generates a set of traffic participant feature attributes based on the spatiotemporal information of the traffic participants, and finally uploads the set of traffic participant feature attributes to the cloud platform.

[0137] For example, the edge computing unit receives the second information, determines the spatiotemporal information of traffic participants through the second information, generates a set of traffic participant feature attributes based on the spatiotemporal information of traffic participants, and finally uploads the set of traffic participant feature attributes to the cloud platform.

[0138] For example, the cloud platform receives the first information, determines the spatiotemporal information of traffic participants through the first information, and generates a set of traffic participant feature attributes based on the spatiotemporal information of traffic participants.

[0139] For example, the cloud platform receives the second information, uses the first information to determine the spatiotemporal information of traffic participants, and generates a set of traffic participant feature attributes based on the spatiotemporal information of traffic participants.

[0140] For example, relying on a cloud platform to receive first and second information, determining the spatiotemporal information of traffic participants through the first and second information, and generating a set of traffic participant feature attributes based on the spatiotemporal information of traffic participants.

[0141] For example, the edge computing unit receives the first information and the second information, determines the spatiotemporal information of the traffic participants through the first information and the second information, generates a set of traffic participant feature attributes based on the spatiotemporal information of the traffic participants, and finally uploads the set of traffic participant feature attributes to the cloud platform.

[0142] To ensure all data acquisition devices support wireless transmission technologies (such as 4G / 5G, Wi-Fi) and achieve real-time or near real-time data transmission back to the central data center, taking camera-sensed image data as an example, ideally, one frame should be saved every 2 seconds to reduce network pressure, data transmission pressure, and avoid frame loss.

[0143] In one embodiment, the first information, the second information, or the traffic participant feature attribute set are transmitted back to the cloud. The transmission method is selected based on the data volume. The data transmission method is determined by the data volume, which influences the data flow on the system. The strategy is determined by the edge computing unit or the server. For example, based on the on-site situation and the data volume, in response to the amount of data acquired / generated (first information, second information, or traffic participant feature attribute set), the edge computing unit determines to select one or more of the following methods: network transmission tool transmission, cloud storage transmission, hard disk transmission, and frame extraction and compression transmission.

[0144] The network transmission tool returns data as follows: if there is an external network on site and the data collection volume is less than 20G, the main time consumption is in network transmission, which generally takes about 1 hour, depending on the amount of data collected.

[0145] The data transfer to the cloud drive is approximately 20-50GB. It is usually done via Baidu Cloud Drive and takes about one day, depending on the amount of data collected.

[0146] For data transfers exceeding 50GB, hard drive transfer is generally used, typically taking 2-3 days. Avoid using FTP for data transfers, as packet loss frequently occurs, rendering the downloaded data unusable.

[0147] The frame extraction and compression return process involves extracting and merging data at specified frame intervals using code, and then returning the data.

[0148] When terminal devices, sensing devices, edge computing devices, and servers exchange data through storage devices, the data storage method is determined by the data volume, which in turn determines the data flow on the system. The strategy is determined by the edge computing unit or the server. For example, based on the data scale or access requirements, the storage method may be at least one of the following: cloud storage, edge storage, solid-state drive storage, distributed file system, and data storage.

[0149] In one embodiment, in response to the amount of data of the first information, second information, or traffic participant feature attribute set acquired / generated, the edge computing unit determines the data storage method as any of the following: cloud storage, edge storage, solid-state drive storage, distributed file system, and data storage.

[0150] In another embodiment, in response to the amount of data of the first information, second information, or traffic participant feature attribute set acquired / generated, the cloud platform server determines the data storage method to be any one of the following: cloud storage, edge storage, solid-state drive storage, distributed file system, and data storage.

[0151] Cloud storage solutions—public, private, or hybrid cloud storage solutions—offer virtually unlimited storage capacity and elastic scalability, making them suitable for handling large-scale, distributed data. Cloud storage also facilitates data backup, disaster recovery, and remote access.

[0152] Edge storage: Data is stored close to the source of the data (such as traffic monitoring cameras), reducing the need for network bandwidth and suitable for real-time analysis or data processing in low-network environments.

[0153] Solid-state drive (SSD) storage: Compared to traditional hard disk drives (HDDs), SSDs offer faster read and write speeds and are suitable for scenarios that require rapid response and processing, such as real-time data analysis.

[0154] The distributed file systems and database storage, such as Hadoop HDFS, Ceph, or Mycat-based distributed database systems, are suitable for handling petabyte-scale data storage and analysis, and are applicable to both structured and unstructured data.

[0155] The choice of storage device or technology depends on the specific application scenario, data processing needs, and budget. For example, if high-performance real-time analysis is required, SSDs or in-memory databases can be used; if data backup and long-term archiving are the focus, cloud storage or tape libraries may be chosen.

[0156] Furthermore, a high-value data storage pool is established to store various types of data. Storage solutions are selected based on data scale and access requirements, including appropriate database systems (such as relational databases, NoSQL databases, and time-series databases). Multiple methods, such as cloud storage, edge storage, and solid-state drives, are employed to ensure efficient data access and compliant management, meeting the needs of different data scales and access requirements. Additionally, data security is ensured during transmission and storage, complying with data protection regulations and anonymizing sensitive information.

[0157] To implement the methods of the various embodiments of this application, in combination with Figures 1-3 This application provides a roadside data mining system, comprising:

[0158] Sensing devices are used to acquire first information and second information.

[0159] Information transmission equipment, i.e. Figure 1The illustrated information and communication (ICT) network device is used to transmit acquired first information and / or second information to a data processing device. In one embodiment, the ICT network includes at least one of the following: V2X, personal mobile communication, and data communication network.

[0160] A data processing device is used to determine the spatiotemporal information of traffic participants in the first and / or second information; it is also used to generate a set of feature attributes of traffic participants. The data processing, i.e., determining the spatiotemporal information of traffic participants in the first and / or second information, can occur on a server, or on an edge computing unit, or it can be preprocessed on the edge computing unit and then sent to the server for further processing.

[0161] In one embodiment, the data processing device includes a multi-source data processing module that processes multiple data sources and aggregates the information. For example, a camera captures a license plate number, point cloud data determines the location, and vehicle information provides satellite positioning.

[0162] It should be noted that this applies to the present application. Figure 1 In the system embodiments of the application scenarios shown, the execution subject of each step of the method provided in the above embodiments of this application can be the same device, or the method can be executed by different devices. For example, the execution subject of steps 210 and 220 can be device 1, and the execution subject of step 230 can be device 2; or, for example, the execution subject of step 210 can be device 1, and the execution subject of steps 220 and 230 can be device 2; and so on.

[0163] In one embodiment, the data processing device includes at least one server 130-1 and an edge computing unit 130-2. The data processing device loads software operation modules containing the functions of steps 210-230, 210-11-13, 210-21-23, and steps 240-250 of the embodiments of this application.

[0164] Furthermore, in one embodiment, the data processing device includes a first data processing device as an edge computing unit and a second data processing device as a server.

[0165] The edge computing unit, located at the roadside node, is used to determine the spatiotemporal information of traffic participants in the first and / or second information; it is also used to generate a set of feature attributes of traffic participants. The functions implemented are as described in steps 210-230 of the embodiments of this application.

[0166] The server, located in the cloud, is used to acquire high-precision map data; it is also used to process data based on the characteristic attribute set of traffic participants and the high-precision map data to generate twin dynamic displays, traffic flow prediction, or accident early warning data. The functions implemented are as described in steps 240-250 of the embodiments of this application.

[0167] The edge computing unit and / or cloud platform exchange traffic participant feature attribute sets through network-side interface ③. The traffic participant feature attribute set is the core of this application and can be defined at the output port of the edge computing unit, or at the input or output port of at least one server.

[0168] The server may be one or more. This application does not impose further limitations.

[0169] Furthermore, it also includes a display terminal 160; the display terminal is used to realize traffic management visualization. It should be noted that the visualization here is a simulated graphic based on high-precision map data, such as represented by vector graphics or other graphic data, containing geometric information and attribute information. The high-precision map is retrieved through an edge computing unit or a server cloud platform. For example, at least one server obtains high-precision map data through the data transmission interface ④ of the high-precision map database, and after data fusion processing, in one embodiment, data processing is performed based on the traffic participant feature attribute set and the high-precision map data to generate twin dynamic display, traffic flow prediction, and accident early warning data in the high-precision map. At least one server sends the twin dynamic display, traffic flow prediction, or accident early warning data to the display terminal through the traffic management data interface ⑤.

[0170] Furthermore, continuous monitoring and evaluation are conducted, a feedback mechanism is established, and the effectiveness of data mining is regularly assessed. Data collection and analysis strategies are adjusted based on actual results. In this embodiment, the application effect is fed back to the data collection and analysis stage, and data collection strategies are adjusted and model algorithms optimized based on the evaluation results. By regularly evaluating the effectiveness of data application, including indicators such as improved traffic efficiency and reduced accident rates, a data-driven continuous improvement mechanism is formed, ensuring the efficient operation of the data closed loop. In one embodiment of this application, the server includes an evaluation module that, based on traffic flow and traffic accident monitoring data, determines data representing changes in traffic efficiency and / or accident rates, and then determines the scope for obtaining first and / or second information based on these changes. The scope includes at least one of the following: the scope of the target area, the time range of data collection, and the range of types and numbers of traffic participants.

[0171] Furthermore, this application optimizes traffic management by applying the analysis results to signal control optimization and road reconstruction and expansion planning to improve overall traffic efficiency; and provides public services: publishing real-time traffic information, providing optimal travel suggestions, and improving the public's travel experience. To achieve these objectives, the traffic management data processing results are fed back to the terminal for application. For the application of traffic management data in terminal devices, terminal status adjustment information is provided to the vehicle terminal to manage the position, speed, and direction of movement of traffic participants, as well as providing information for adjusting vehicle operating status, driving behavior, and communication between traffic participants.

[0172] For example:

[0173] For the vehicle-mounted terminal, the server or edge computing device feeds back traffic management instruction information through a second information interface. This traffic management instruction information includes the vehicle-mounted terminal's identification information, which is associated with the identification of traffic participants (i.e., vehicles), and traffic management data within a target area related to the traffic participant's identification. This data may include at least one of the following: twin dynamic display, traffic flow prediction, or accident warning data; or a storage address containing at least one of these data. The vehicle-mounted terminal obtains this traffic management instruction information through an application, for example, by real-time transmission or retrieval of storage devices, and then visualizes it on the vehicle-mounted terminal's graphical user interface. Further, the application generates vehicle-mounted terminal status adjustment information, which includes at least one of the following: new indication information representing the vehicle's position, speed, direction, or operating status.

[0174] For personal mobile terminals, the server or edge computing device feeds back traffic management instruction information through a second information interface. This traffic management instruction information includes the personal mobile terminal's identification information, which is associated with the identification (name or organization name) of a traffic participant (i.e., a pedestrian or driver), and traffic management data within a target area related to the traffic participant's identification. This data may include at least one of the following: twin dynamic display, traffic flow prediction, or accident warning data; or a storage address containing at least one of these data. The personal mobile terminal obtains this traffic management instruction information through an application, for example, by real-time transmission or retrieval from a storage device, and then visualizes it on the graphical user interface of the in-vehicle terminal. Further, the application generates personal terminal status adjustment information, which includes at least one of the following: new instruction information indicating the location, speed, or direction of the pedestrian or driver. This instruction information can be displayed on a graphical application interface and output via voice through a traffic control or navigation application, providing information on the planned location, speed, and direction of movement to form a planned route.

[0175] For traffic control terminals, the server or edge computing device feeds back traffic management instruction information through a second information interface. This traffic management instruction information includes the identification information of the traffic control terminal, which is associated with road traffic control facilities, and traffic management data within a target area related to the road traffic facilities. For example, it includes at least one type of data such as twin dynamic display, traffic flow prediction, or accident warning data, or a storage address containing at least one type of data. The traffic control terminal obtains this traffic management instruction information through an application, for example, by real-time transmission or retrieval of storage devices. This data is then visualized on the traffic control terminal's graphical user interface or displayed on the traffic control facilities. Further, the application generates traffic control terminal status adjustment information, which includes at least one of the following: information indicating the position, speed, or direction of traffic participants.

[0176] The roadside data mining solution proposed in this application includes methods for data acquisition, data recycling and transmission, data processing and analysis, high-value data mining, and how to build a smart transportation system based on traffic data. This roadside data mining solution can not only significantly improve the precision of urban traffic management, but also promote environmental protection and energy conservation, and enhance emergency response capabilities.

[0177] 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.

[0178] Therefore, this application also proposes a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the methods described in any embodiment of this application.

[0179] 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.

[0180] 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.

[0181] 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.

[0182] Figure 7This is a structural diagram of a computer device according to an embodiment of this application. Furthermore, this application also proposes an electronic device (or computing device) including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method described in any embodiment of this application.

[0183] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory. Memory may include non-persistent storage in computer-readable media, random access memory (RAM), and / or non-volatile memory such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media. Computer-readable media includes both permanent and non-persistent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information that can be accessed by the computing device. As defined in this article, computer-readable media do not include transient media, such as modulated data signals and carrier waves.

[0184] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0185] Those skilled in the art will understand that, unless otherwise stated, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be understood that when a device or component is “connected” to another device or component, it may be directly connected to the other device or component, or there may be an intermediary device or component. Furthermore, the term “connection” as used herein may include partially wireless connections as well as partially wired connections.

[0186] In the description of this application, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances. Furthermore, in the description of this application, unless otherwise stated, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0187] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A roadside data mining method, characterized in that, Including the following steps: Acquire first information and second information; the first information includes roadside sensing information acquired through roadside sensing devices; the second information includes terminal status information acquired through information and communication networks. Determine the spatiotemporal information of traffic participants in the first and / or second information; A set of traffic participant feature attributes is generated, which includes the number, type, and spatiotemporal information of traffic participants, as well as data representing traffic events determined based on the above information.

2. The roadside data mining method according to claim 1, characterized in that, It also includes the following steps: Acquire high-precision map data; Data processing is performed based on the characteristic attribute set of traffic participants and high-precision map data to generate twin dynamic display, traffic flow prediction or accident early warning data.

3. The roadside data mining method according to claim 1, wherein the information and communication network comprises at least one of the following: V2X, personal mobile communication, and data communication network; the terminal comprises at least one of the following: vehicle-mounted terminal, traffic control terminal, and personal mobile terminal; characterized in that, The vehicle terminal status information includes at least one of the following: information indicating the vehicle's position, speed, direction, or operating status; Traffic control terminal status information includes at least one of the following: information indicating the position, speed, or direction of traffic participants; Personal mobile terminal status information includes at least one of the following: information indicating the location, speed, or direction of a pedestrian or driver.

4. The roadside data mining method according to claim 1, characterized in that, Data processing of the first information and / or the second information includes at least one of the following: Data cleaning, feature extraction, pattern recognition, outlier handling, data augmentation, or processing model optimization.

5. The roadside data mining method according to claim 1, characterized in that, Based on the data scale or access requirements, the storage method is determined to be at least one of the following: cloud storage, edge storage, solid-state drive storage, distributed file system, or data storage.

6. The roadside data mining method according to claim 1, characterized in that, The first information, the second information, or the set of traffic participant characteristic attributes are transmitted back to the cloud. The transmission methods include: transmission via network transmission tools, transmission via cloud storage, transmission via hard disk, or transmission via frame extraction and compression.

7. A roadside data mining system, characterized in that, The method for implementing the roadside data mining method according to any one of claims 1-7 comprises: a sensing device, an information transmission device, and a data processing device; Sensing devices are used to acquire first information and second information; The information transmission device is used to transmit the acquired first information and second information to the data processing device; The data processing device is used to determine the spatiotemporal information of traffic participants in the first information and / or the second information; it is also used to generate a set of feature attributes of traffic participants.

8. The roadside data mining system according to claim 7, characterized in that, The data processing device includes a first data processing device as an edge computing unit and / or a second data processing device as a server.

9. The roadside data mining system according to claim 7, characterized in that, The data processing device includes a server and an edge computing unit; The edge computing unit, located at the roadside node, is used to determine the spatiotemporal information of traffic participants in the first information and / or the second information; it is also used to generate a set of feature attributes of traffic participants. The server, located in the cloud, is used to acquire high-precision map data; it is also used to process data based on the characteristic attribute set of traffic participants and high-precision map data to generate twin dynamic display, traffic flow prediction or accident early warning data.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-6.

11. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-6.

Images