A method for dividing urban expressway weaving areas based on aerial video trajectory extraction

Abstract

The present invention discloses a method for dividing an urban expressway interweaving area based on aerial video trajectory extraction, including collecting static parameters of the expressway interweaving area based on a drone, and extracting vehicle trajectory data using video tracking technology; classifying vehicle lane-changing behaviors, collecting lane-changing information of vehicles in the expressway interweaving area and obtaining coordinate data when changing lanes; using a K-means clustering algorithm to divide the expressway interweaving area according to the coordinate data when changing lanes; performing collision risk assessment on the divided areas of the expressway interweaving area according to the vehicle trajectory data, and drawing a risk distribution map; and making traffic flow optimization suggestions for the expressway interweaving area according to the risk distribution map. While improving the traffic flow efficiency of urban intersections and highway intersections, the present invention provides strong support for the intelligentization and sustainable development of urban transportation systems.

Description

Technical Field

[0001] The present invention belongs to the technical field of traffic management of expressway interweaving areas, and in particular relates to a method for dividing urban expressway interweaving areas based on aerial video trajectory extraction. Background Art

[0002] With the continuous increase of urban traffic flow, traffic management in highway intersecting areas has become increasingly complex. The accident causal chain theory believes that accidents are the result of a series of chain events. Intervention in traffic conflict links may help change the distribution of traffic conflict severity, thereby reducing the incidence and severity of accidents and improving the operational safety level of intersecting areas.

[0003] Based on this, the first prior art is a traffic control method for an interweaving area based on the spatial distribution characteristics of traffic conflicts: the method uses a traffic conflict spatial distribution prediction model to analyze the traffic flow conditions in the interweaving area under different traffic parameters and makes corresponding traffic control measures, which include:

[0004] Step 1: Use the traffic parameters of the urban expressway weaving area to build a VISSIM simulation model and use the model to obtain conflict data.

[0005] Step 2: Input the conflict data into the K-means clustering algorithm, classify them according to the location where the traffic conflicts occur, and divide the urban expressway interweaving area into different location areas.

[0006] Step 3: Analyze the number of traffic conflicts of different degrees in each location area under lane change conflicts and rear-end collisions. Change the traffic parameters of the interweaving area of ​​the urban expressway and input them into the VISSIM simulation model to obtain a larger sample size; calculate the traffic risk index K of each location area in each sample, and use all K values ​​to draw the cumulative frequency curve; divide the values ​​of 20%, 40%, 60% and 80% in the cumulative frequency curve into different safety levels, corresponding to fairly safe, relatively safe, critically safe and dangerous respectively; divide the safety level into two road conditions of 0 and 1, where the road condition K′ is defined as 0 when the K value is below critical safety, otherwise the road condition K′ is defined as 1.

[0007] Step 4: Construct a traffic risk index K. Enter the traffic parameters under different weaving zone operating conditions into the traffic conflict spatial distribution prediction model to obtain the road condition K′ value. If the K′ values ​​in all location areas are 0, the current road condition is determined to be safe.

[0008] Step 5: Construct a logistic regression classification model, use the data set simulated by the VISSIM simulation model to train the logistic regression classification model, and use the trained logistic regression classification model as a traffic conflict spatial distribution prediction model.

[0009] Step 6: Use the traffic conflict spatial distribution prediction model to analyze the traffic flow conditions in the weaving area under different traffic parameters and make corresponding traffic control measures.

[0010] The disadvantages of this prior art are:

[0011] (1) The weaving zone division method adopted by this scheme only classifies the areas according to the location where the conflict occurs, but ignores the spatial complexity of the lanes and does not consider the specific lane change type that causes the conflict. This simple division method is not reasonable in nature.

[0012] (2) When analyzing conflicts in different regions, the scheme simply records the number of conflicts without considering the spatiality within the region. Although this method can provide a certain degree of conflict statistics, it ignores the possible spatial relationships and spatial distribution patterns between different locations within the region.

[0013] (3) The data used in this scheme comes from the simulation model. Although the simulation process tries to simulate the real-world situation as much as possible, it is still difficult to completely match the real-world data.

[0014] Prior art 2: A method for evaluating the safety of an expressway weaving area based on traffic conflicts: The present invention discloses a method for evaluating the safety of an expressway weaving area based on traffic conflicts, which can realize the safety evaluation of the expressway weaving area in sections, effectively identify the weak points of the safety of the expressway weaving area, and then improve the traffic safety of the expressway weaving area in a targeted manner, which specifically includes:

[0015] Step 1: Divide the expressway weaving area into several sections according to traffic operation characteristics; divide the weaving area into merging area, internal area of ​​the weaving area and diversion area.

[0016] Step 2: Collect traffic data from each section, and use a drone equipped with a camera to record video of the interweaving area of ​​the expressway; collect traffic conflicts in the interweaving area and record the sections where traffic conflicts occur; use video tracking software to draw a trajectory map of the conflicting vehicles, and divide traffic conflicts into flanking conflicts and rear-end conflicts; and divide traffic conflicts into general conflicts and severe conflicts based on the collision time TTC of the two vehicles.

[0017] Step 3: Select the traffic conflict risk index of each section as a safety evaluation index, wherein the safety evaluation index includes different types of traffic conflict risk indexes and traffic conflict risk indexes of different severity levels;

[0018] Step 4: According to the safety evaluation index values ​​of each section obtained in step 3, the gray clustering analysis method is used to determine the safety level of each section in the weaving area according to the evaluation matrix results. Select different types of traffic conflict risk index K1 and traffic conflict risk index K2 of different severity for each section to establish an evaluation matrix; divide the safety level of the weaving section into four levels, which are four gray classes, and determine the whitening value and whitening weight function under different gray classes respectively; determine the clustering weight values ​​of different safety evaluation indicators under different safety levels; according to the evaluation matrix results, select the level with the largest evaluation value as the safety evaluation level.

[0019] The disadvantages of this prior art are:

[0020] (1) The division of the expressway weaving area in this plan is based on its traffic operation characteristics, and only considers static factors such as the road's geometric dimensions and lane functions, but ignores the impact of dynamic factors such as actual vehicle behavior and peak hours and traffic conflicts.

[0021] (2) The data set of this scheme is small and is not suitable for the grey cluster analysis method. The safety level of each section in the weaving area obtained has large errors. Summary of the invention

[0022] The purpose of the present invention is to provide a method for dividing urban expressway weaving areas based on aerial video trajectory extraction in view of the above-mentioned deficiencies in the prior art, so as to solve the problem that the existing expressway weaving area division has great limitations.

[0023] In order to achieve the above object, the technical solution adopted by the present invention is:

[0024] A method for dividing an urban expressway interweaving area based on aerial video trajectory extraction comprises the following steps:

[0025] S1. Collect static parameters of the expressway weaving area based on drones and extract vehicle trajectory data using video tracking technology;

[0026] S2. Classify the lane-changing behaviors of vehicles, collect lane-changing information of vehicles in the interweaving area of ​​the expressway, and obtain the coordinate data of the lane-changing;

[0027] S3, based on the coordinate data during lane change, the K-means clustering algorithm is used to divide the expressway interweaving area;

[0028] S4. Based on the vehicle trajectory data, a collision risk assessment is performed on the area divided into the expressway interweaving area, and a risk distribution map is drawn;

[0029] S5. Based on the risk distribution map, make recommendations on optimizing traffic flow in the expressway weaving area.

[0030] Furthermore, in step S1, the static parameters of the expressway interweaving area are collected based on the drone, including:

[0031] A drone is used to record video of the expressway weaving area, and static parameters of the expressway weaving area are obtained based on the recorded video. The static parameters include: weaving area width, lane width and weaving area length.

[0032] Furthermore, the vehicle trajectory data in step S1 includes: the vehicle center of mass coordinate point (x E ,y E ), and the four vertex coordinates of the vehicle frame outline

[0033] Furthermore, the vehicle lane changing behavior is classified in step S2, including:

[0034] The lane-changing behaviors are divided into merging, merging and overtaking.

[0035] The merge type is defined as the left lane change behavior of merging into the main lane on the merging ramp;

[0036] The merging type is defined as the behavior of changing lanes to the right by making advance preparations from the main lane to the target diverging ramp;

[0037] Overtaking is defined as the act of changing lanes to the left to overtake the vehicle ahead while driving in the main lane.

[0038] Furthermore, the lane changing information of the vehicle in the expressway weaving area in step S2 includes: the lane changing type, position and direction of the vehicle.

[0039] Furthermore, the coordinate data of the lane change is obtained in step S2, including: the coordinate point where the center of mass of the vehicle passes through the curb line when changing lanes is the lane change coordinate point, and the distance between the lane change coordinate point and the starting point of the expressway interweaving area is x n , the fitting function of the curb line is f(x n ), then the lane change coordinate point is (x n ,f(x n )).

[0040] Further, step S3 includes:

[0041] S3.1. For several lane change coordinate points (x n ,f(x n )) are marked, and the marked lane-changing coordinate points are used as samples, and k lane-changing coordinate points are randomly selected as the center points;

[0042] S3.2, calculating the distances between the n*k lane change coordinate points and the k center points;

[0043] S3.3, assign the sample to the cluster corresponding to the nearest center point;

[0044] S3.4. Calculate the number of samples and the center point in each cluster grouping;

[0045] S3.5, iterate S3.2 to S3.4 until the clustering algorithm converges;

[0046] S3.6. Based on the converged clustering, the expressway weaving area is divided into the merging area, lane changing area and overtaking area.

[0047] Furthermore, in step S4, collision risk assessment is performed on the area divided into the expressway interweaving area according to the vehicle trajectory data, including:

[0048]

[0049]

[0050] Among them, T E (t) is the collision time of the following vehicle E; d y is the shortest distance between the leading vehicle and the following vehicle E; v Fy is the lateral speed of the leading vehicle F; The ordinate of a vertex in the vehicle frame outline of the following vehicle E; The ordinate of a vertex in the vehicle box outline of the leading vehicle F;

[0051] If T E When (t)<1s, it is considered that there is a serious collision risk between the leading vehicle F and the following vehicle E;

[0052] When T E When (t) is 1 to 2 s, it is considered that there is a greater risk of collision between the leading vehicle F and the following vehicle E;

[0053] When T E When (t)>2s, it is considered that there is a general collision risk between the leading vehicle F and the following vehicle E.

[0054] The method for dividing the urban expressway weaving area based on aerial video trajectory extraction provided by the present invention has the following beneficial effects:

[0055] 1. The present invention analyzes real vehicle trajectory data, considers lane functionality and vehicle lane change information, and fully considers spatial heterogeneity, so as to more accurately and effectively divide the interweaving areas of urban expressways; through risk assessment of the severity of conflicts within the space, a more targeted optimization solution is provided for traffic management, which can alleviate traffic congestion and improve traffic efficiency; and while improving the traffic flow efficiency of urban intersections and highway intersections, the present invention provides strong support for the intelligent and sustainable development of urban transportation systems.

[0056] 2. The present invention takes into account lane functionality and vehicle lane-changing information, and divides the weaving area of ​​urban expressways. By deeply studying the actual driving conditions of vehicles, it is possible to more accurately understand the dynamic characteristics within the weaving area and divide it into more effective and targeted areas.

[0057] 3. The present invention conducts a comprehensive risk assessment of the complexity of the internal space of the interweaving area of ​​urban expressways, and fully considers the spatial heterogeneity. Since there are differences in road functions, traffic flow, vehicle speed, etc. between different areas, these differences directly affect the traffic safety risks of the interweaving area. Therefore, in the risk assessment process, the present invention not only considers global factors, but also focuses on analyzing the characteristics of each area in order to more accurately identify and solve potential safety hazards.

[0058] 4. Dynamic area division method considering spatiality: In this process, the motives for vehicle lane changes are analyzed and screened, and then targeted vehicle lane change types are labeled and data is classified; then, based on this classification logic, the algorithm will draw a distribution map of lane changing behaviors in the interweaving area, which serves as the basis for dynamic area division; through this method, the lane changing risk factors are fully spatially quantified, and on this basis, the complexity of spatial heterogeneity is integrated, providing a more refined decision support system for subsequent traffic flow planning and safety strategy formulation.

[0059] 5. In-depth integration of TTC theory: In the process of analyzing the internal severity of each area, the horizontal and vertical dimensions of TTC theory are integrated. The extension of this theory not only overcomes the limitations of the traditional TTC model in vehicle lane change risk assessment, but also adds consideration of the vehicle's lateral dynamic movement and its relative speed changes; such advancement enables the present invention to more accurately depict and evaluate the collision risk caused by lane changing behavior in three-dimensional space, thereby improving the predictability of traffic safety and the accuracy of management. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 The present invention is a flow chart of a method for dividing an urban expressway interweaving area based on aerial video trajectory extraction.

[0061] Figure 2 It is a schematic diagram of static parameters of the interleaving area of ​​the present invention.

[0062] Figure 3 Schematic diagram of the coordinates of the lane-changing point of the vehicle according to the present invention.

[0063] Figure 4 This is a schematic diagram of the collection results of the input point of the present invention.

[0064] Figure 5This is a schematic diagram of the collection results of the export point of the present invention.

[0065] Figure 6 This is a schematic diagram of the overtaking point collection results of the present invention.

[0066] Figure 7 Schematic diagram of vehicle vector distance and speed according to the present invention. DETAILED DESCRIPTION

[0067] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0068] Example 1

[0069] This embodiment provides a method for dividing the urban expressway weaving area based on aerial video trajectory extraction. It starts from real vehicle trajectory data and divides the urban expressway weaving area by analyzing the severity and number of traffic conflicts in different areas. It can more accurately understand the dynamic characteristics inside the weaving area and divide it into more effective and targeted areas. Figure 1 , which specifically includes the following steps:

[0070] Step S1, collecting static parameters of the expressway weaving area based on the drone, and extracting vehicle trajectory data using video tracking technology, which specifically includes the following contents:

[0071] The data set collected in this embodiment mainly comes from real data extracted from urban expressways during the morning and evening peak hours. These data have higher credibility and can better reflect the real situation of urban traffic conditions;

[0072] Specifically, this embodiment uses a drone to record a video of the expressway weaving area, and obtains static parameters of the expressway weaving area based on the recorded video, wherein reference Figure 2 ,Static parameters include: weaving area width, lane width and weaving area length.

[0073] At the same time, this embodiment uses video tracking technology to collect vehicle trajectory data and obtain the vehicle center of mass coordinate point (x E ,y E ), select the four vertices of the vehicle outline

[0074] Finally, the collected trajectory data is preprocessed, including removing outliers and smoothing trajectory information, to ensure the accuracy and reliability of the data.

[0075] Step S2: classify the lane-changing behavior of the vehicle, collect the lane-changing information of the vehicle in the expressway interweaving area and obtain the coordinate data when the lane is changed, which specifically includes the following contents:

[0076] Classify vehicle lane-changing behaviors, including merging, merging, and overtaking;

[0077] The merge type is defined as the behavior of changing lanes to the left on the merging ramp to merge into the main lane;

[0078] The merging type is defined as the behavior of changing lanes to the right by making advance preparations from the main lane to the target diverging ramp;

[0079] Overtaking is defined as the act of changing lanes to the left to overtake the vehicle ahead while driving in the main lane.

[0080] Collect lane change information in the weaving area, including the type, position, and direction of lane changes by the vehicle, and record the coordinates of the location where the lane change occurs.

[0081] When the vehicle changes lanes, the coordinate point where the center of mass passes through the curb line is the lane change coordinate point. Figure 3 , let the distance between the lane change coordinate point of the nth sample and the starting point of the expressway weaving area be x n , the fitting function of the curb line is f(x n )(f(x n )=1, 2, 3), then the lane change coordinate point is (x n ,f(x n )).

[0082] Step S3: Based on the coordinate data during lane change, the K-means clustering algorithm is used to divide the expressway interweaving area into regions, which specifically includes the following steps:

[0083] Step S3.1: for a number of lane change coordinate points (x n ,f(x n )) for marking, reference Figure 4 , 5 , 6, and the marked lane change coordinate points are taken as samples, denoted as (x n ,f(x n )), arbitrarily select k lane-changing coordinate points as the center points;

[0084] Step S3.2, calculating the distances between the n*k lane change coordinate points and the k center points;

[0085] Step S3.3, assign the sample to the cluster corresponding to the nearest center point;

[0086] Step S3.4, calculating the number of samples and the center point in each cluster grouping;

[0087] Step S3.5, iterate S3.2 to S3.4 until the clustering algorithm converges;

[0088] Step S3.6: Based on the converged clustering, the expressway weaving area is divided into the merging area, the lane changing area and the overtaking area.

[0089] Step S4: Based on the vehicle trajectory data, a collision risk assessment is performed on the area divided into the expressway interweaving area, and a risk distribution map is drawn, which specifically includes the following contents:

[0090] The TTC calculation formula of the following vehicle E relative to the leading vehicle F at time t is:

[0091]

[0092] Among them, T E (t) is the collision time of the following vehicle E, in seconds; G E (t) is the distance in front of the following vehicle E, in meters; v(t) is the speed of the vehicle, in kilometers per hour; the classic TTC model usually describes vehicle behavior as a one-dimensional problem, so the analysis of lane-changing conflicts in weaving areas needs to be further improved, mainly considering lateral conflicts, referring to Figure 7 , showing the change in lateral distance between vehicle F and the following vehicle E when the vehicle F in the merging area changes lanes to the left.

[0093] refer to Figure 7 ,θ E v is the tilt angle of the leading vehicle F when changing lanes to the left; Fy is the lateral speed of the leading vehicle F, in km / h; v Fx is the longitudinal speed of the leading vehicle F, in km / h; d y is the shortest distance between the leading vehicle and the following vehicle E, generally the distance from the nearest vertex to the facing end of vehicle F. The improved TTC calculation formula of this embodiment is:

[0094]

[0095]

[0096] Among them, T E (t) is the collision time of the following vehicle E; The ordinate of a vertex in the vehicle frame outline of the following vehicle E; The ordinate of a vertex in the vehicle box outline of the leading vehicle F;

[0097] If T E When (t)<1s, it is considered that there is a serious collision risk between the leading vehicle F and the following vehicle E;

[0098] When T E When (t) is 1 to 2 s, it is considered that there is a greater risk of collision between the leading vehicle F and the following vehicle E;

[0099] When T E When (t)>2s, it is considered that there is a general collision risk between the leading vehicle F and the following vehicle E.

[0100] Subsequently, based on the above judgment, a risk distribution map is drawn.

[0101] Step S5: Based on the risk distribution map, make traffic flow optimization suggestions for the expressway weaving area to improve the traffic efficiency of each area.

[0102] Although the specific implementation of the invention is described in detail in conjunction with the drawings, it should not be understood as limiting the scope of protection of this patent. Within the scope described in the claims, various modifications and variations that can be made by those skilled in the art without creative work still fall within the scope of protection of this patent.

Claims

1. A method for dividing urban expressway weaving areas based on aerial video trajectory extraction, characterized in that: The following steps are involved: S1. Collect static parameters of the expressway weaving area based on drones and extract vehicle trajectory data using video tracking technology; S2. Classify the lane-changing behaviors of vehicles, collect lane-changing information of vehicles in the interweaving area of ​​the expressway, and obtain the coordinate data of the lane-changing; S3. Based on the coordinate data during lane change, the K-means clustering algorithm is used to divide the expressway weaving area into regions; and based on the clustering after convergence, the expressway weaving area is divided into a merging area, a lane change area, and an overtaking area; S4. According to the vehicle trajectory data, a collision risk assessment is performed on the area divided into the expressway interweaving area, and a risk distribution map is drawn; S5. Providing traffic flow optimization suggestions for the expressway interweaving area based on the risk distribution map; The step S1 collects static parameters of the expressway interweaving area based on the drone, including: Using a drone to record a video of the expressway weaving area, and based on the recorded video, obtaining static parameters of the expressway weaving area, the static parameters including: weaving area width, lane width and weaving area length; The step S2 classifies the lane changing behavior of the vehicle, including: The lane-changing behaviors are divided into merging, merging and overtaking. The merge type is defined as a left lane change behavior that merges into the main lane on the merging ramp; The merging type is defined as the behavior of changing lanes to the right by making advance preparations from the main lane to the target diverging ramp; The overtaking type is defined as the behavior of changing lanes to the left to overtake the vehicle ahead while driving in the main lane; The lane change information of the vehicle in the expressway weaving area in step S2 includes: the lane change type, position and direction of the vehicle; In step S4, based on the vehicle trajectory data, a collision risk assessment is performed on the area divided into the expressway interweaving area, including: in, For following vehicles The collision time; The shortest distance between the leading vehicle and the following vehicle E; is the lateral speed of the leading vehicle F; For following vehicles The ordinate of a vertex in the vehicle frame selection outline; The ordinate of a vertex in the vehicle box outline of the leading vehicle F; like <1s, it is considered that the leading vehicle F and the following vehicle There is a serious risk of collision; when When the time is 1~2s, it is considered that the leading vehicle F and the following vehicle There is a greater risk of collision; when >2s, the leading vehicle F and the following vehicle There is a general collision risk in between.

2. The method for dividing the urban expressway weaving area based on aerial video trajectory extraction according to claim 1 is characterized in that: The vehicle trajectory data in step S1 includes: the vehicle centroid coordinate point , and the four vertex coordinates of the vehicle's frame outline , , , .

3. The method for dividing the urban expressway weaving area based on aerial video trajectory extraction according to claim 1 is characterized in that: The coordinate data of the lane change in step S2 includes: the coordinate point where the center of mass of the vehicle passes through the curb line when changing lanes is the lane change coordinate point, and the distance between the lane change coordinate point and the starting point of the expressway interweaving area is , the fitting function of the curb line is , then the lane change coordinate point is .

4. The method for dividing the urban expressway weaving area based on aerial video trajectory extraction according to claim 3 is characterized in that: The step S3 comprises: S3.

1. For several lane change coordinate points of different types Marking is performed, and the marked lane-changing coordinate points are used as samples, and k lane-changing coordinate points are randomly selected as the center points; S3.2, calculating the distances between the n*k lane change coordinate points and the k center points; S3.3, assign the sample to the cluster corresponding to the nearest center point; S3.

4. Calculate the number of samples and the center point in each cluster grouping; S3.5, iterate S3.2~S3.4 until the clustering algorithm converges; S3.

6. Based on the converged clustering, the expressway weaving area is divided into the merging area, lane changing area and overtaking area.

Images