Vehicle collision risk association mode assessment method based on directed weighted graph

Abstract

The invention belongs to the technical field of driving risk evaluation, and discloses a directed weighted graph-based vehicle collision risk association mode evaluation method, which comprises the following steps of: 1, identifying vehicles passing through a road section of an operation area from a traffic image video and extracting information through a computer vision technology, the extraction information comprises the position, the speed and the distance of each vehicle; 2, the extracted information is used for quantifying the collision risk ITA of a target vehicle # imgabs0 # and an interaction object # imgabs1 #, a vehicle risk quantification matrix is formed, the target vehicle # imgabs2 # is a rear vehicle, and the interaction object # imgabs3 # is a front vehicle and an isolation facility; and step 3, introducing a directed multiple weighted graph, depicting risk association features of the vehicle group, and achieving the purpose of visually and quantitatively analyzing the collision risk relationship between all vehicles and isolation facilities in the operation area.

Description

Technical Field

[0001] This application belongs to the technical field of driving risk assessment, and specifically relates to a method for evaluating the vehicle collision risk association pattern based on a directed weighted graph. Background Art

[0002] Highway work areas are specific areas specially set to effectively ensure the safety of vehicles during construction operations. Their settings mainly involve daily maintenance operations and renovation and expansion projects, and the demands for these two aspects of projects in China are both increasing continuously. On the one hand, many sections of China's highways have entered the mid-late stage of their service life, and the frequency of work area settings will continue to rise. On the other hand, the continuous growth of traffic volume will not only increase the frequency of daily maintenance operations but also require the improvement of the bearing capacity of highways through renovation and expansion projects. To ensure regional traffic access, the renovation and expansion of highways often adopt the mode of "constructing while maintaining traffic flow". However, the setting of work areas has changed the existing traffic environment, thus significantly increasing the risk of vehicle collisions within this section of the road.

[0003] Within the work area, the interactions between different types of traffic participants (such as vehicles, traffic cones, construction vehicles, etc.) are complex and changeable, which poses challenges to the accurate quantification of collision risks and the characterization of the situation. Due to limitations in theoretical research and data collection methods, existing research often ignores these multi-object interactions, resulting in a series of problems such as incomplete risk assessment, unclear optimization objects for static safety measures, and low timeliness of dynamic early warning. Therefore, it is necessary to study how to accurately quantify and evaluate the collision risks between vehicles and different traffic participants from the perspectives of potential field theory and computer vision technology in the complex scenario of multi-object interactions in work areas. This has extremely important practical significance for significantly improving the traffic safety level of work area sections.

[0004] Based on the above defects of the existing technology, there is an urgent need to propose a method for evaluating the vehicle collision risk association pattern in work areas based on graph theory and directed weighted graphs to solve the above problems. Summary of the Invention

[0005] The purpose of this application is to solve the problems of the existing technology, and provides a method for evaluating the vehicle collision risk association pattern based on a directed weighted graph, providing a theoretical basis for evaluating the collision risks between vehicles and different traffic participants in such sections to ensure the driving safety of work area sections.

[0006] To solve the technical problems, the technical solution of this application is: A method for evaluating the vehicle collision risk association pattern based on a directed weighted graph, including the following steps: Step 1: Through computer vision technology, identify the vehicles passing through the work area section from traffic image videos and extract information, where the extracted information includes the positions, speeds, and spacings of each vehicle; Step 2: Quantify the target vehicle by using the extracted information and the interaction object 's collision risk ITA to form a vehicle risk quantification matrix, where the target vehicle is a rear vehicle, and the interaction object is a front vehicle and a separation facility; Step 3: Introduce a directed multi-weighted graph to characterize the risk association characteristics of the vehicle group; Step 3-1: Introduce a directed multi-weighted graph, and use the target vehicle passing through the work area and the interaction object as the node set of the directed multi-weighted graph, denoted as ; Step 3-2: Take the interaction relationship between the target vehicle and the interaction object as the directed edge of the directed multi-weighted graph, denoted as , where the directed edge refers to the direction from the target vehicle to the interaction object; Step 3-3: Take the collision risk between the target vehicle and the interaction object ITA as the weight of each directed edge, and denote it as the weight of the directed edge ; Step 3-4: Use the directed multi-weighted graph to represent the collision risk relationship between all target vehicles in the work area and the interaction object

[0007] to obtain a vehicle collision risk system correlation characterization graph. ITA ) ITA This application uses the improved time to accident (

[0008] Preferably, the specific content of Step 2 is: According to the position, speed and spacing of the vehicle, calculate the collision risk between the target vehicle and the interaction object ITA to construct a vehicle risk quantification matrix; The calculation formula of the collision risk ITA is: ; Wherein: and are the position and speed of the target vehicle at t moment; , and are the position, speed and spacing of the interaction object at t moment; is the correction coefficient.

[0009] Preferably, the isolation facility is a traffic cone, a guardrail or a vehicle parked abnormally.

[0010] Preferably, step 3 is specifically as follows: Step 3-1: Introduce a directed multi-weighted graph, and use all the vehicles and isolation facilities passing through the work area as the node set of the directed multi-weighted graph, denoted as ; Step 3-2: Take the interaction relationships between vehicles and between vehicles and isolation facilities as the directed edges of the directed multi-weighted graph, denoted as , and the directed edge refers to the target vehicle pointing to the interaction object ; Step 3-3: Take the collision risks ITA between vehicles and between vehicles and isolation facilities as the weights of each directed edge, and denote them as the weights of the directed edge ; Step 3-4: Use the directed multi-weighted graph to represent the collision risk relationship between all target vehicles in the work area and the interaction object to obtain a vehicle collision risk system relevance characterization graph.

[0011] Preferably, when there are two isolation facilities, the vehicle collision risk system relevance characterization graph has two centers, that is, each vehicle node has at least two directed edges, and all passing vehicles point to these two centers, which is called the multi-center type; When there is one isolation facility, the vehicle collision risk system relevance characterization graph has one center, which is called the single-center type; When the isolation facility is not considered, or high-risk vehicles are screened out, the vehicle collision risk system relevance characterization graph is multiple groups, which is called the non-center type.

[0012] Preferably, after obtaining the correlation representation diagram of the vehicle collision risk system, according to the risk threshold of the extreme field force F V screen the high-risk vehicles, retain the nodes corresponding to the high-risk vehicles and the directed edges between the high-risk vehicles, realize the correlation representation of the high-risk vehicles, and the screened correlation representation diagram of the vehicle collision risk system is divided into multiple groups.

[0013] Preferably, the F V is used to quantify the collision risk of the target vehicle, and the calculation formula of the extreme field force F V is: ; In the formula: F V is the extreme field force; is the collision potential field strength; is the field force between the target vehicle and the interaction object ; is the risk quantity of the interaction object .

[0014] Preferably, according to the extreme field force F V and taking the vehicle behind as the target vehicle , determine the following-following and overtaking phenomena between vehicles and whether there is a high-risk interaction relationship and whether it is a high-risk vehicle through different types of dual-node or triple-node correlation models. The dual-node correlation model includes dual-node type A and dual-node type B, and the triple-node correlation model includes triple-node type A, triple-node type B, triple-node type C, triple-node type D, triple-node type E, and triple-node type F.

[0015] Compared with the prior art, the advantages of this application are: (1) In response to the need for vehicle group collision risk assessment, this application introduces a directed multi-weighted graph and proposes a vehicle collision risk correlation pattern assessment method. This method visually shows the risk correlation between all vehicles and isolation facilities in the work area section using a directed multi-weighted graph, and screens high-risk vehicles according to the risk threshold of the extreme field force F V , so as to quickly identify high-risk vehicle groups, clarify high-risk interactions, and further improve the accuracy of risk quantification and the comprehensiveness of risk assessment in the work area; (2) The evaluation method of this application combines a directed multi-weighted graph, which not only shows the directional relationship between nodes, but also expresses the strength or cost of these relationships through the weights of directed edges. When using a directed multi-weighted graph to characterize the relevance system of vehicle collision risks in the work area, due to different isolation facilities, the relevance system characterization graph of collision risks shows different styles, which can reflect the impact of different isolation facilities on the relevance of the collision risk system; (3) This application visually displays the relevance between vehicles through a directed multi-weighted graph, which can effectively quantify the interaction risk relationship between vehicles and isolation facilities in the work area section, clarify high-risk interactions, improve the accuracy of the high-risk vehicle group identification method, and thus provide real-time and predictive support for the safety management of the work area. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flowchart of the vehicle collision risk association mode evaluation method based on a directed weighted graph of this application; Figure 2 is a relevance system characterization graph of vehicle collision risks in Embodiment 1 of this application; Figure 3 is a relevance system characterization graph of vehicle collision risks of the multi-center type in Embodiment 1 of this application; Figure 4 is a relevance system characterization graph of vehicle collision risks of the single-center type in Embodiment 1 of this application; Figure 5 is a relevance system characterization graph of vehicle collision risks of the non-center type in Embodiment 1 of this application; Figure 6 is a relevance system characterization graph of vehicle collision risks of several types of high-risk vehicles in Embodiment 1 of this application; Figure 7 and Figure 9 is a relevance system characterization graph of vehicle collision risks in the upstream transition section during operation in Embodiment 2 of this application; Figure 8 and Figure 10 is a relevance system characterization graph of vehicle collision risks during non-operation in Embodiment 2 of this application; Figure 11 is a high-risk interaction relationship graph between high-risk vehicles in Embodiment 2 of this application; Figure 12 is a high-risk interaction relationship graph between high-risk vehicles in Embodiment 2 of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] The present application will be described in detail below in conjunction with the accompanying drawings and specific embodiments, but the present application is not limited to these embodiments. The present application covers any alternatives, modifications, equivalent methods, and solutions made within the essence and scope of the present application. For the public to have a thorough understanding of the present application, specific details are described in detail in the following embodiments of the present application, but those skilled in the art can fully understand the present application without the description of these details.

[0018] The present application discloses a method for evaluating the vehicle collision risk association pattern based on a directed weighted graph, including the following steps: Step 1: Through computer vision technology, identify the vehicles passing through the work area section from the traffic image video and extract information, including the position, speed, and spacing of each vehicle. Step 2: Utilize the extracted information to quantify the collision risk between the target vehicle ITA and the interaction object to form a vehicle risk quantification matrix, where the target vehicle is the rear vehicle, and the interaction object is the front vehicle and the isolation facility. Step 3: Introduce a directed multi-weighted graph to characterize the risk association characteristics of the vehicle group. Step 3-1: Introduce a directed multi-weighted graph, and use the target vehicle passing through the work area and the interaction object as the node set of the directed multi-weighted graph, denoted as ; Step 3-2: Take the interaction relationship between the target vehicle and the interaction object as the directed edge of the directed multi-weighted graph, denoted as , where the directed edge refers to the direction from the target vehicle to the interaction object; Step 3-3: Take the collision risk between the target vehicle and the interaction object ITA as the weight of each directed edge, and denote it as the weight of the directed edge ; Step 3-4: Use the directed multi-weighted graph to represent the collision risk relationship between all target vehicles and interaction objects in the work area, and obtain the vehicle collision risk system correlation characterization graph.

[0019] Preferably, the specific steps of step 2 are as follows: According to the position, speed and spacing of the vehicle (the units of the position, speed and spacing of the vehicle use conventional units and just keep the correspondence), calculate the target vehicle and the interaction object the collision risk ITA between them, and construct a vehicle risk quantification matrix; The collision risk ITA (s -1 ) is calculated by the formula: ; In the formula: and are the position (m) and speed (m / s) of the target vehicle at t time; , and are the position (m), speed (m / s) and spacing (m) of the interaction object at t time; is the correction coefficient; The units of the above parameters use conventional units and just keep the correspondence.

[0020] Preferably, the isolation facility is a traffic cone, a guardrail or a vehicle parked abnormally.

[0021] Preferably, the specific steps of step 3 are as follows: Step 3-1: Introduce a directed multi-weighted graph. All vehicles and isolation facilities passing through the work area are used as the node set of the directed multi-weighted graph, denoted as ; Step 3-2: Use the interaction relationships between vehicles and between vehicles and isolation facilities as the directed edges of the directed multi-weighted graph, denoted as , and the directed edge refers to the target vehicle pointing to the interaction object ; Step 3-3: Use the collision risks ITA between vehicles and between vehicles and isolation facilities as the weights of each directed edge, and denote them as the weights of the directed edge ; Step 3-4: Use the directed multi-weighted graph to represent the collision risk relationships between all target vehicles in the work area and the interaction object , and obtain a vehicle collision risk system correlation characterization graph.

[0022] Preferably, when there are two isolation facilities, the correlation representation graph of the vehicle collision risk system has two centers, that is, each vehicle node has at least two directed edges, and all passing vehicles point to these two centers, which is called the multi-center type; When there is one isolation facility, the correlation representation graph of the vehicle collision risk system has one center, which is called the single-center type; When the isolation facility is not considered, or high-risk vehicles are screened out, the correlation representation graph of the vehicle collision risk system is multiple groups, which is called the non-center type.

[0023] Preferably, after obtaining the correlation representation graph of the vehicle collision risk system, according to the extreme field force F V The risk threshold of is used to screen high-risk vehicles, retain the nodes corresponding to high-risk vehicles and the directed edges between high-risk vehicles, realize the correlation representation of high-risk vehicles, and the correlation representation graph of the vehicle collision risk system after screening is divided into multiple groups.

[0024] Preferably, the extreme field force F V The calculation formula of is: ; In the formula: F V Is the extreme field force; Is the collision potential field strength; Is the target vehicle And the interaction object The field force between; Is the interaction object The risk amount of; The units of the above parameters use conventional units and just keep corresponding.

[0025] Preferably, according to the extreme field force F V And taking the vehicle behind as the target vehicle , different types of the two-node or three-node correlation model are used to determine the following and overtaking phenomena between vehicles and whether there is a high-risk interaction relationship between vehicles and whether they are high-risk vehicles. The two-node correlation model includes two-node type A and two-node type B, and the three-node correlation model includes three-node type A, three-node type B, three-node type C, three-node type D, three-node type E, and three-node type F.

[0026] Embodiment 1 Such as Figure 1As shown in the figure, the present application provides a method for evaluating the vehicle collision risk association pattern based on a directed weighted graph, which includes the following steps: Step 1: Through computer vision technology, identify the vehicles passing through the work area section from the traffic image video and extract information, where the extracted information includes the position, speed, and spacing of each vehicle; Step 2: Utilize the extracted information to quantify the collision risk between the target vehicle and the interaction object ITA , forming a vehicle risk quantification matrix, where the target vehicle is the rear vehicle, and the interaction object is the front vehicle and the isolation facility, and the isolation facility is a traffic cone; Step 3: Introduce a directed multi-weighted graph to depict the risk association characteristics of the vehicle group; Introduce a directed multi-weighted graph, with vehicles and traffic cones as the node set of the directed multi-weighted graph, denoted as ; the interaction relationships between vehicles and between vehicles and traffic cones are used as the directed edges of the graph, denoted as , and the directed edge refers to the one pointing from the target vehicle to the interaction object ; the collision risk ITA between vehicles is used as the weight of each edge, and is denoted as the weight of the directed edge .

[0027] Use the directed multi-weighted graph to represent the collision risk relationships between all vehicles in the work area; for the purpose of being able to visually and quantitatively analyze the collision risk relationships between all vehicles and traffic cones in the work area; as Figure 2 shown, it is a diagram for characterizing the system relevance of vehicle collision risks in the work area based on a directed multi-weighted graph.

[0028] As Figures 3 - 5 shown, it is a diagram for characterizing the system relevance of three typical vehicle collision risks; As Figure 3 shown, it is a diagram for characterizing the system relevance of the vehicle collision risk of the multi-center type. When considering multiple isolation facilities (such as traffic cones, guardrails, and abnormally parked vehicles) simultaneously, the diagram for characterizing the system relevance of the vehicle collision risk has two centers, that is, each vehicle node has at least two edges, and all vehicles passing through the work area point to these two centers, which is called the multi-center type; As Figure 4As shown in the figure, it is a correlation representation diagram of the vehicle collision risk system of the single - center type. When only considering an object fixed in the road area space, the correlation representation diagram of the vehicle collision risk system has only one center, which is called the single - center type; Such as Figure 5 shown in the figure, it is a correlation representation diagram of the vehicle collision risk system of the non - center type. When not considering the isolation facilities or screening out high - risk vehicles, the correlation representation diagram of the vehicle collision risk system shows multiple groups, which is called the non - center type.

[0029] Furthermore, on the basis of representing the correlation of the vehicle collision risk system in the directed multi - weighted graph operation area, further screen the high - risk vehicles according to the risk threshold of the extreme field force F V That is, only retain the nodes corresponding to high - risk vehicles and the directed edges between high - risk vehicles, and the correlation representation of high - risk vehicles can be realized. The screened directed multi - weighted Figure 1 is generally divided into multiple groups.

[0030] Such as Figure 6 shown in the figure, it is a correlation representation diagram of the vehicle collision risk system of several types of high - risk vehicles. According to the extreme field force F V and taking the vehicle behind as the target vehicle , it is possible to determine the following - following, overtaking phenomena between vehicles and whether there is a high - risk interaction relationship between vehicles through different types of double - node or triple - node correlation models, Figure 6 In which (a) is double - node type A, (b) is double - node type B, (c) is triple - node type A, (d) is triple - node type B, (e) is triple - node type C, (f) is triple - node type D, (g) is triple - node type E, and (h) is triple - node type F.

[0031] Embodiment 2 To make the purpose, technical solution and advantages of the present application clearer, Embodiment 2 of the present application will explain the specific implementation cases of the present application with the collision risk relationships of specific cases AW, AN, BW, and BN and in combination with the accompanying drawings.

[0032] Such as Figures 7 - 10 shown in the figure, it is a correlation representation diagram of the vehicle collision risk system in four cases. It should be noted that due to considering the limitation of the available space in the figure, the weights in the figure are omitted, and in this embodiment, only one type of isolation facility, the traffic cone, is considered, and all traffic cones are regarded as a linear entity.

[0033] Figure 7 And Figure 9 shown in the figure is the correlation representation diagram of the vehicle collision risk system in the upstream transition section during the operation period ( Figure 7 represents case AW, Figure 9Representative case BW), whose style is of the single-center type, with the center being the node representing the traffic cone.

[0034] Figure 8 and Figure 10 The correlation representation diagram of the vehicle collision risk system during non-operation shown in Figure 8 Representative case AN, Figure 10 Representative case BN), whose style is of the non-center type.

[0035] It should be noted that since the isolation facilities are not considered, many vehicles passing alone are not shown in the figure. At the same time, it is found from the figure that there are vehicle groups with multiple complex collision risk correlation relationships in the four cases, and the number of vehicle groups during operation is significantly higher than that during non-operation. In addition, by comparing Figure 8 and Figure 10 it can be found that: the complexity of the collision risk relationship and the number of vehicle groups on section B during non-operation are significantly higher than those on section A, which means that generally the collision risk on section B may be higher than that on section A, and the setting of the operation area increases this trend.

[0036] As Figures 11 - 12 shown, it is the high-risk interaction relationship diagram between high-risk vehicles.

[0037] Based on the correlation representation of the vehicle collision risk system in the above four cases, high-risk vehicles are screened out, that is, only the edges and nodes related to high-risk vehicles are retained to identify high-risk vehicle groups and evaluate the interaction relationships between high-risk vehicles. Figure 11 Samples A and B in Figure 12 represent two relatively complex high-risk vehicle association model samples extracted from the surveillance video,

[0038] It can be seen from this that the high-risk interaction relationship with only two nodes is not easily affected by other vehicles, while the collision risks between cluster vehicles with more than two nodes are interrelated and are more likely to be affected by the improper risk avoidance behavior of one of the vehicles, resulting in accidents.

[0039] According to statistics, there are a total of 17 vehicle groups with high-risk interaction relationships in case AW, among which 15 groups are two-node association models and 2 groups are complex association models with three or more nodes, accounting for 11.76%; there are 31 vehicle groups with high-risk interaction relationships in case BW, among which 19 groups are two-node association models and 12 groups are complex association models with three or more nodes, accounting for 38.71%. This shows that the collision risk correlation relationship in case BW is significantly higher than that in case AW, which also means that for the same form of work area setting on different sections of the same highway, there may be significant differences in the impact on road traffic safety.

[0040] In summary, an evaluation method for vehicle collision risk association patterns based on a directed weighted graph provided by an embodiment of the present application introduces a directed multi-weighted graph, using the vehicles passing through the work area and traffic cones as the node set of the directed multi-weighted graph, and the interaction relationships between vehicles and between vehicles and traffic cones as the directed edges of the directed multi-weighted graph; quantifies the collision risk between the target vehicle and the interacting vehicle through the extreme field force of the work area collision risk potential field, and then uses the directed multi-weighted graph to represent the collision risk relationships between all vehicles in the work area. Ultimately, it aims to visually and quantitatively analyze the collision risk relationships between all vehicles and traffic cones in the work area. It can quickly identify high-risk vehicle groups, clarify high-risk interaction relationships, and further improve the accuracy of risk quantification and the comprehensiveness of risk assessment in the work area.

[0041] The preferred embodiments of the present application are described in detail above, but the present application is not limited to the above embodiments. Within the knowledge scope of those of ordinary skill in the art, various changes can be made without departing from the purpose of the present application.

[0042] Many other changes and modifications can be made without departing from the concept and scope of the present application. It should be understood that the present application is not limited to specific embodiments, and the scope of the present application is defined by the appended claims.

Claims

1. A method for evaluating the vehicle collision risk association pattern based on a directed weighted graph, characterized in that It includes the following steps: Step 1: Through computer vision technology, identify the vehicles passing through the work area section from traffic image videos and extract information, where the extracted information includes the positions, speeds and spacings of each vehicle; Step 2: Quantify the target vehicle using the extracted information and the interaction object 's collision risk ITA to form a vehicle risk quantification matrix, where the target vehicle is the rear vehicle, and the interaction object is the front vehicle and the isolation facility; Step 3: Introduce a directed multi-weighted graph to characterize the risk association characteristics of the vehicle group; Step 3-1: Introduce a directed multi-weighted graph, with the target vehicle passing through the work area and the interaction object as the node set of the directed multi-weighted graph, denoted as ; Step 3-2: The target vehicle and the interaction object The interaction relationship between them is used as the directed edge of the directed multi-weighted graph, denoted as , The directed edge refers to the one directed from the target vehicle to the interaction object ; Step 3-3: The collision risk between the target vehicle and the interaction object ITA is used as the weight of each directed edge, denoted as the weight of the directed edge ; Step 3-4: Use a directed multi-weighted graph to represent all target vehicles in the work area and the interaction objects to obtain a graph representing the correlation of the vehicle collision risk system 2. The vehicle collision risk association pattern evaluation method based on a directed weighted graph according to claim 1, wherein The specific content of step 2 is as follows: Calculate the target vehicle based on the position, speed and spacing of the vehicle and the interaction object for the collision risk ITA , and construct a vehicle risk quantification matrix; The collision risk ITA The calculation formula is as follows: ； In the formula: and is the position and speed of the target vehicle at t the moment; , and are the position, speed and spacing of the interaction object at t moment; is the correction coefficient.

3. The vehicle collision risk association pattern evaluation method based on a directed weighted graph according to claim 1, wherein The isolation facility is a traffic cone, a guardrail or a vehicle parked abnormally.

4. The vehicle collision risk association pattern evaluation method based on a directed weighted graph according to claim 3, wherein The specific content of Step 3 is: Step 3-1: Introduce a directed multi-weighted graph. All vehicles and isolation facilities passing through the work area are used as the node set of the directed multi-weighted graph, denoted as ; Step 3-2: Take the interaction relationships between vehicles and between vehicles and isolation facilities as directed edges of a directed multi-weighted graph, denoted as , and the directed edge refers to the target vehicle pointing to the interaction object ; Step 3-3: Take the collision risks between vehicles and between vehicles and isolation facilities ITA as the weights of each directed edge, and denote them as the weights of the directed edge ; Step 3-4: Using a directed multi-weighted graph to represent all target vehicles in the work area and the interaction objects to obtain a correlation representation graph of the vehicle collision risk system.

5. The method for evaluating the vehicle collision risk association pattern based on a directed weighted graph according to claim 4, wherein: When there are two isolation facilities, the vehicle collision risk system association characterization graph has two centers, that is, each vehicle node has at least two directed edges, and all passing vehicles point to these two centers, which is called the multi-center type; When there is one isolation facility, the vehicle collision risk system association characterization graph has one center, which is called the single-center type; When the isolation facility is not considered, or high-risk vehicles are screened out, the vehicle collision risk system association characterization graph is multiple groups, which is called the non-center type.

6. The vehicle collision risk association pattern evaluation method based on a directed weighted graph according to claim 5, wherein: After obtaining the correlation representation diagram of the vehicle collision risk system, according to the risk threshold of the extreme field force F V screen the high-risk vehicles, retain the nodes corresponding to the high-risk vehicles and the directed edges between the high-risk vehicles, realize the correlation representation of the high-risk vehicles, and the correlation representation diagram of the vehicle collision risk system after screening is divided into multiple groups.

7. The vehicle collision risk correlation pattern evaluation method based on a directed weighted graph according to claim 6, characterized in that: The extreme field force F V The calculation formula is as follows: ； In the formula: F V is the extreme field force; is the field strength of the collision potential field; is the target vehicle and the interaction object the field force therebetween; For an interaction object The amount of risk 8. The vehicle collision risk association pattern evaluation method based on a directed weighted graph according to claim 6, characterized in that: According to the extreme field force F V and taking the vehicle behind as the target vehicle , determine the following-following and overtaking phenomena between vehicles and whether there is a high-risk interaction relationship and whether it is a high-risk vehicle among vehicles through different types of dual-node or triple-node association models. The dual-node association model includes dual-node Type A and dual-node Type B, and the triple-node association model includes triple-node Type A, triple-node Type B, triple-node Type C, triple-node Type D, triple-node Type E, and triple-node Type F.

Images