Vehicle travel trajectory reconstruction method considering the topological structure of the checkpoint network

Abstract

In view of the limitations of the prior art, the present invention proposes a method for reconstructing vehicle travel trajectories that takes into account the topological structure of the checkpoint network; the present invention takes into account the traffic rules of the road network nodes, and uses the checkpoint network topological structure and the preset travel time threshold to divide two consecutive single trips, which can quickly and accurately identify the travel trajectories of vehicles with different travel purposes, greatly reducing the amount of calculation, and can be applied to large-scale data sets; combining the constructed checkpoint network topological structure and the Dijkstra algorithm to restore the vehicle travel trajectory, and incorporating the spatial topological attributes of the road network into the research process of trajectory reconstruction, so that the trajectory reconstruction has a high degree of environmental adaptability. The trajectory reconstruction result of the present invention has a high degree of restoration and performs well in actual thousand-kilometer-level urban road network traffic scenarios.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle trajectory reconstruction in intelligent transportation, and in particular to a vehicle travel trajectory reconstruction method considering a checkpoint network topology structure. Background Art

[0002] Vehicle travel paths contain a large number of individual travel characteristics and patterns, which can be regarded as static storage of road network traffic status information. By extracting and analyzing the travel trajectories of all vehicles traveling in the road network, the actual traffic operation conditions of the entire road network can be restored. As shown in the Chinese invention patent with an announcement date of 2021.01.08: A method and system for reconstructing the spatiotemporal trajectory of commuting vehicles based on license plate recognition data, which aims to utilize the traffic distribution characteristics of commuting vehicles, complete the missing spatial position of the trajectory, improve the accuracy of restoring time nodes, and accurately know the time when commuting vehicles appear at each intersection in the road network.

[0003] Due to the differences in parameters such as road network coverage, sampling frequency, and sampling time among traffic information collection equipment, and the frequent occurrence of data anomalies and missing data during data collection and transmission, it is necessary to further reconstruct the vehicle travel trajectory to improve the accuracy of trajectory recognition results.

[0004] According to whether the detection equipment is fixed, the trajectory-oriented traffic information collection technology is generally divided into two methods: fixed-point detection and mobile detection. Fixed-point detection equipment includes induction coils, infrared detectors, etc.; mobile detection equipment includes floating vehicles, GPS, mobile phone signaling, etc. At present, most vehicle travel trajectory reconstruction methods based on GPS data have a relatively single data source and cannot be applied to large-scale road network-level trajectory extraction. Moreover, as the road network coverage rate of traffic checkpoint equipment gradually increases, traffic checkpoint data with full coverage, wide application and easy acquisition has become an effective basis for vehicle travel trajectory reconstruction research. In the existing vehicle travel trajectory reconstruction technology, the idea of ​​single trip division and optimal trajectory decision is generally used to reconstruct the trajectory. Among them, the division of vehicle single trips is mainly based on the behavioral conditions of single or multiple travelers such as travel time, driving distance, speed, etc., without considering the spatial structure constraints of the actual road network, and there are many calculation steps. At the same time, in the existing optimal trajectory decision technology, the TOPSIS decision method, traffic flow theory and other ideas are usually used to estimate the actual operation trajectory of the vehicle. The data processing process is relatively complicated and is only applicable to a small-scale road network. Moreover, it only considers the static road network structure and ignores the traffic rules of the actual road network nodes. There are certain errors, which is difficult to be applied in practice for large-scale data sets and urban road networks. Summary of the invention

[0005] In view of the limitations of the prior art, the present invention proposes a vehicle travel trajectory reconstruction method considering the topological structure of the checkpoint network. The technical solution adopted by the present invention is:

[0006] A vehicle travel trajectory reconstruction method considering the topological structure of a checkpoint network includes the following steps:

[0007] S1, obtaining original card slot data, adjusting the format of the original card slot data, and filtering out invalid data;

[0008] S2, generating a bicycle travel information set with a time-related order according to the processing result of step S1;

[0009] S3, constructing a checkpoint network topology structure based on the bicycle travel information set and the actual traffic rules of the road network nodes; the checkpoint network topology structure is a weighted directed graph describing the spatiotemporal relationship between the road network checkpoints, and is used to provide a geographic reference for subsequent address matching and trajectory restoration operations;

[0010] S4, dividing a single trip trajectory according to the bicycle travel information set and a preset travel time threshold;

[0011] S5, performing address matching and trajectory restoration according to the checkpoint network topology structure and a single travel trajectory in combination with the Dijkstra algorithm.

[0012] Compared with the prior art, the present invention takes into account the traffic rules of the road network nodes, uses the checkpoint network topology structure and the preset travel time threshold to divide two consecutive single trips, can quickly and accurately identify the travel trajectories of vehicles with different travel purposes, greatly reduces the amount of calculation, and can be applied to large-scale data sets; combines the constructed checkpoint network topology structure and Dijkstra algorithm to restore the vehicle travel trajectory, and incorporates the road network space topology attributes into the research process of trajectory reconstruction, so that the trajectory reconstruction has a higher environmental adaptability. The trajectory reconstruction result of the present invention has a high degree of restoration and performs well in actual thousand-kilometer urban road network traffic scenes.

[0013] As a preferred solution, in step S1:

[0014] Obtain OSM city road network maps and traffic checkpoint equipment distribution maps, and collect original checkpoint data detected by road network checkpoint equipment within a specific time period;

[0015] According to the preset format standard, the attributes of the original checkpoint data including the checkpoint number, the date of vehicle passing and the time of vehicle passing are formatted; the checkpoint passing records in which the vehicle license plate or the checkpoint number is not identified in the original checkpoint data, as well as the duplicate vehicle passing records are filtered out.

[0016] Furthermore, in step S2, "one week" is used as a statistical time node to determine the single-vehicle travel information of the vehicle within one week:

[0017] The processing result of step S1 is divided into groups identified by vehicle license plates, each group including all vehicle passing records at different times;

[0018] Each group is sorted in ascending order according to the detected passing time, and a bicycle travel information set CR = {θ1, θ2, …, θ N},θ i =(b i ,d i ,t i ),i=1,2,…,N; where N is the number of times the vehicle is detected in the checkpoint data within one week, θ i is a record of a vehicle passing through a checkpoint, b i is the bayonet number corresponding to the bayonet device, d i is the date when the vehicle passes by the camera device, t i is the corresponding passing time, and t1 <t2<…<t N .

[0019] Furthermore, in the step S3, the following process is included:

[0020] According to the bicycle travel information set and in combination with the actual traffic rules of the road network nodes, high-frequency checkpoint pairs that are adjacent in time are extracted as adjacent checkpoint pairs, and the average travel time of vehicles passing through the adjacent checkpoint pairs is calculated;

[0021] An adjacency matrix is ​​constructed according to the adjacent bayonet pairs.

[0022] Furthermore, in step S3, high-frequency checkpoint pairs that are adjacent in time are extracted as adjacent checkpoint pairs in the following manner, and the average travel time of vehicles passing through the adjacent checkpoint pairs is calculated:

[0023] According to the passing time of each vehicle passing record in the bicycle travel information set CR, all time-related checkpoint pairs are identified (b i ,b j ), that is, the bayonet numbers of adjacent bayonet ports in the set are connected in sequence;

[0024] Taking the preset minimum support SupSUM as the threshold, the card port pairs (b i ,b j ) frequency of occurrence f i,j If f i,j ≥SupSUM, then the bayonet pair is retained as an adjacent bayonet pair;

[0025] Calculate the travel time of all vehicles passing through adjacent checkpoint pairs

[0026] The 55% travel time on the road segment is selected as the adjacent checkpoint pair b i , j The average travel time between i,j ,Right now in, Vehicle k passes through checkpoint b j The time when the vehicle passes is detected. Vehicle k passes through checkpoint b i The time of the vehicle passing detected at , T (i,j) For passing through adjacent bayonet pairs (b i ,b j ) is the set of travel times of all vehicles, that is, n is the number of vehicles passing through the checkpoint pair within a specific time period.

[0027] Furthermore, in step S3, an adjacency matrix is ​​constructed in the following manner to mark the extracted adjacent checkpoint pairs as outliers:

[0028] Extract all the adjacent bayonet pairs contained in the bayonet {b1, b2, …, b m} as the row and column labels of the adjacency matrix, m is the number of checkpoints contained in the list of adjacent checkpoint pairs, then the size of the adjacency matrix is ​​m×m;

[0029] According to the preset red light time threshold R (R∈(0,20]), 70% to 90% of the road network travel time is selected as the travel time threshold B; at i,j ≤R and at i,j >B's adjacent bayonet pair (b i ,b j ) is marked as an abnormal bayonet pair;

[0030] Determine the adjacency matrix weight ω i,j The value of i,j >B, then it means bayonet b i , b j Not directly connected, so i,j =0; if at i,j ≤B, then ω i,j =at i,j .

[0031] Furthermore, in step S4, the single trip trajectory is divided by repeating the following steps:

[0032] Traverse the bicycle travel information set CR={θ1,θ2,…,θ N},θ i =(b i ,d i ,t i ),i=1,2,…,N;

[0033] Compare the vehicles passing through the checkpoint (b i ,b i+1 ) at the time of passing the vehicle i d i+1 Whether they are consistent: If the dates are different and the vehicle travels through the checkpoint pair This means that the vehicle is at the checkpoint (b i ,b i+1 ) There is a long period of static state between the two checkpoints, that is, the travel trajectories before and after the checkpoint pair do not belong to the same single trip;

[0034] According to the adjacent bayonet pair (b i ,b i+1 ) between the average travel time at i,i+1 And the marked abnormal checkpoints are used to determine the vehicle operation status: if at i,i+1 ≤R, it means that the checkpoint pair does not conform to the distance that the actual vehicle can move at normal speed or is a duplicate record caused by the waiting time for the red light at the intersection. Therefore, only the adjacent checkpoint pairs are retained (b i ,b i+1 ) in the next checkpoint, i.e. checkpoint b i+1 ; if at i,j >B, indicating that the vehicle is stationary for a long time between the adjacent checkpoints;

[0035] The adjacent checkpoint pairs (b i ,b i+1 ) to generate two consecutive single trips, namely SR1 = {b1, b2, …, b i}, SR2 = {b i+1 ,b i+2 ,…,b N}.

[0036] Furthermore, in step S5, the following process is included:

[0037] According to the row and column labels of the adjacency matrix of the checkpoint network topology structure and the geographical location information of the checkpoint device, the longitude and latitude data of the checkpoint vertex corresponding to the checkpoint network topology structure in the actual road network are queried; if the corresponding longitude and latitude data cannot be queried, the checkpoint identified by the checkpoint number is identified as an invalid checkpoint; the invalid checkpoints and their corresponding association function relationships in the checkpoint network topology structure and its adjacency matrix are deleted;

[0038] Traverse the single trip trajectory SR in order to find the invalid checkpoints in the single trip trajectory. According to the position of the invalid checkpoints in the travel trajectory, they are divided into the following two cases for processing:

[0039] If the invalid checkpoint is a single trip trajectory SR = {b1, b2, ..., b i-1 ,b i ,b i+1 ,…,b D} starting point b1 or end point b D , then directly delete the invalid card port;

[0040] If the invalid checkpoint is the travel trajectory SR={b1,b2,…,b i-1 ,b i ,b i+1 ,…,b D}Intermediate bayonet b i , use Dijkstra algorithm to find the card port b on the card port network topology structure i-1 and b i+1 The shortest path between If the shortest path exists, the invalid checkpoints in the travel trajectory are replaced with the intermediate nodes of the shortest path in sequence; otherwise, the invalid checkpoints are directly deleted.

[0041] The present invention also provides the following contents:

[0042] A computer storage medium stores a computer program, which, when executed by a processor, implements the steps of the aforementioned vehicle travel trajectory reconstruction method considering the topological structure of a checkpoint network.

[0043] A computer device includes a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor. When the computer program is executed by the processor, the steps of the vehicle travel trajectory reconstruction method considering the checkpoint network topology structure in Example 1 are implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of the steps of a vehicle travel trajectory reconstruction method considering a checkpoint network topology structure provided by an embodiment of the present invention;

[0045] Figure 2 It is a schematic diagram of the principle of dividing a single travel trajectory in an embodiment of the present invention;

[0046] Figure 3 It is a schematic diagram of the principle of the address matching and track restoration process in an embodiment of the present invention;

[0047] Figure 4 This is an example of a distribution map of urban road networks and traffic checkpoint equipment used in an embodiment of the present invention;

[0048] Figure 5 This is an example of the bayonet network topology and adjacency matrix constructed in an embodiment of the present invention.

[0049] Figure 6 This is an example of an incomplete travel trajectory diagram of a vehicle in an embodiment of the present invention.

[0050] Figure 7 This is an example of a travel trajectory diagram reconstructed in an embodiment of the present invention.

[0051] Figure 8 It is the coincidence rate between the reconstructed trajectory and the GPS data in the embodiment of the present invention. DETAILED DESCRIPTION

[0052] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0053] It should be clear that the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the embodiments of the present application.

[0054] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the embodiments of the present application. The singular forms of "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0055] When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims. In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are only used to distinguish similar objects, and do not have to be used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.

[0056] In addition, in the description of this application, unless otherwise specified, "multiple" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0057] In order to solve the limitations of the prior art, this embodiment provides a technical solution, and the technical solution of the present invention is further described below in conjunction with the drawings and embodiments.

[0058] Example 1

[0059] Please refer to Figure 1 ,A vehicle travel trajectory reconstruction method considering the topological structure of the checkpoint network, includes the following steps:

[0060] S1, obtaining original card slot data, adjusting the format of the original card slot data, and filtering out invalid data;

[0061] S2, generating a bicycle travel information set with a time-related order according to the processing result of step S1;

[0062] S3, constructing a checkpoint network topology structure based on the bicycle travel information set and the actual traffic rules of the road network nodes; the checkpoint network topology structure is a weighted directed graph describing the spatiotemporal relationship between the road network checkpoints, and is used to provide a geographic reference for subsequent address matching and trajectory restoration operations;

[0063] S4, dividing a single trip trajectory according to the bicycle travel information set and a preset travel time threshold;

[0064] S5, performing address matching and trajectory restoration according to the checkpoint network topology structure and a single travel trajectory in combination with the Dijkstra algorithm.

[0065] Compared with the prior art, the present invention takes into account the traffic rules of the road network nodes, uses the checkpoint network topology structure and the preset travel time threshold to divide two consecutive single trips, can quickly and accurately identify the travel trajectories of vehicles with different travel purposes, greatly reduces the amount of calculation, and can be applied to large-scale data sets; combines the constructed checkpoint network topology structure and Dijkstra algorithm to restore the vehicle travel trajectory, and incorporates the road network space topology attributes into the research process of trajectory reconstruction, so that the trajectory reconstruction has a higher environmental adaptability. The trajectory reconstruction result of the present invention has a high degree of restoration and performs well in actual thousand-kilometer urban road network traffic scenes.

[0066] As a preferred embodiment, in step S1:

[0067] Obtain OSM city road network maps and traffic checkpoint equipment distribution maps, and collect original checkpoint data detected by road network checkpoint equipment within a specific time period;

[0068] According to the preset format standard, the attributes of the original checkpoint data including the checkpoint number, the date of vehicle passing and the time of vehicle passing are formatted; the checkpoint passing records in which the vehicle license plate or the checkpoint number is not identified in the original checkpoint data, as well as the duplicate vehicle passing records are filtered out.

[0069] Specifically, the original checkpoint data is the passing record of all vehicles passing through the traffic checkpoint equipment on the road network within a certain period of time, including attribute columns such as passing date (GCRQ), passing time (GCSJ), checkpoint number (KKBH), and vehicle license plate (CLTMBH).

[0070] Since traffic network security checkpoint detection equipment will produce certain errors in vehicle identification, data collection and transmission, it may cause abnormal situations such as duplicate checkpoint data records and missing fields, which will cause certain interference to subsequent operations. Therefore, the original checkpoint data that does not conform to the format specifications needs to be checked, updated and modified before use.

[0071] By extracting all the checkpoint passing records of the vehicle within a specific time period and connecting them in the order of the detected passing time, the single vehicle travel information can be obtained. Therefore, further, in the step S2, "one week" is used as the statistical time node to determine the single vehicle travel information of the vehicle within one week:

[0072] The processing result of step S1 is divided into groups identified by vehicle license plates, each group including all vehicle passing records at different times;

[0073] Each group is sorted in ascending order according to the detected passing time, and a bicycle travel information set CR = {θ1, θ2, …, θ N},θ i =(b i ,d i ,t i ),i=1,2,…,N; where N is the number of times the vehicle is detected in the checkpoint data within one week, θ i is a record of a vehicle passing through a checkpoint, b i is the mount number corresponding to the mount device, d i is the date when the vehicle passes by the camera device, t i is the corresponding passing time, and t1 <t2<…<t N .

[0074] Furthermore, in the step S3, the following process is included:

[0075] According to the bicycle travel information set and in combination with the actual traffic rules of the road network nodes, high-frequency checkpoint pairs that are adjacent in time are extracted as adjacent checkpoint pairs, and the average travel time of vehicles passing through the adjacent checkpoint pairs is calculated;

[0076] An adjacency matrix is ​​constructed according to the adjacent bayonet pairs.

[0077] Furthermore, in step S3, high-frequency checkpoint pairs that are adjacent in time are extracted as adjacent checkpoint pairs in the following manner, and the average travel time of vehicles passing through the adjacent checkpoint pairs is calculated:

[0078] According to the passing time of each vehicle passing record in the bicycle travel information set CR, all time-related checkpoint pairs are identified (b i ,b j ), that is, the bayonet numbers of adjacent bayonet ports in the set are connected in sequence;

[0079] Taking the preset minimum support SupSUM as the threshold, the card port pairs (b i ,b j ) frequency of occurrence f i,j If f i,j ≥SupSUM, then the bayonet pair is retained as an adjacent bayonet pair;

[0080] Calculate the travel time of all vehicles passing through adjacent checkpoint pairs

[0081] In order to eliminate the influence of extreme values ​​on the overall situation, the 55% travel time on the road section is selected as the adjacent checkpoint b i , b j The average travel time between i,j ,Right now in, Vehicle k passes through checkpoint b j The time when the vehicle passes is detected. Vehicle k passes through checkpoint b i The time of the vehicle passing detected at , T (i,j) For passing through adjacent bayonet pairs (b i ,b j ) is the set of travel times of all vehicles, that is, n is the number of vehicles passing through the checkpoint pair within a specific time period.

[0082] Furthermore, in order to eliminate the influence of abnormal operation of the checkpoint equipment, two consecutive single trips of the vehicle and the waiting time of the red light at the intersection on the checkpoint network topology, in step S3, an adjacency matrix is ​​constructed in the following way to mark the extracted adjacent checkpoint pairs as outliers:

[0083] Extract all the adjacent bayonet pairs contained in the bayonet {b1, b2, …, b m} as the row and column labels of the adjacency matrix, m is the number of checkpoints contained in the list of adjacent checkpoint pairs, then the size of the adjacency matrix is ​​m×m;

[0084] According to the preset red light time threshold R (R∈(0,20]), 70% to 90% of the road network travel time is selected as the travel time threshold B; at i,j ≤R and at i,j >B's adjacent bayonet pair (b i ,b j ) is marked as an abnormal bayonet pair;

[0085] Determine the adjacency matrix weight ω i,j The value of i,j >B, then it means bayonet b i , b j Not directly connected, so i,j =0; if at i,j ≤B, then ω i,j =at i,j .

[0086] The same vehicle may have multiple consecutive trips in one day, so it is necessary to further reasonably divide single trips in the single-vehicle travel information set. The separation position between two consecutive single trips can be determined based on whether the vehicle is in a long-term static state. Therefore, according to the aforementioned red light time threshold R and travel time threshold B, the running state of the vehicle between the time-related checkpoints is defined as the basis for dividing the single trip trajectory. Therefore, further, in the step S4, the single trip trajectory is divided by repeating the following steps:

[0087] Traverse the bicycle travel information set CR={θ1,θ2,…,θ N},θ i =(b i ,d i ,t i ),i=1,2,…,N;

[0088] Compare the vehicles passing through the checkpoint (b i ,b i+1 ) at the time of passing the vehicle i ,d i+1 Whether they are consistent: If the dates are different and the vehicle travels through the checkpoint pair This means that the vehicle is at the checkpoint (b i ,b i+1 ) There is a long period of static state between the two checkpoints, that is, the travel trajectories before and after the checkpoint pair do not belong to the same single trip;

[0089] According to the adjacent bayonet pair (b i ,b i+1 ) between the average travel time at i,i+1 And the marked abnormal checkpoints are used to determine the vehicle operation status: if at i,i+1 ≤R, it means that the checkpoint pair does not conform to the distance that the actual vehicle can move at normal speed or is a duplicate record caused by the waiting time for the red light at the intersection. Therefore, only the adjacent checkpoint pairs are retained (b i ,b i+1 ) in the next checkpoint, i.e. checkpoint b i+1 ; if at i,j >B, indicating that the vehicle is stationary for a long time between the adjacent checkpoints;

[0090] The adjacent checkpoint pairs (b i ,b i+1 ) to generate two consecutive single trips, namely SR1 = {b1, b2, …, b i}, SR2 = {b i+1 ,b i+2 ,…,b N}.

[0091] Since the checkpoints included in the checkpoint network topology may not completely match the geographic location information of the actual checkpoint equipment, some of the checkpoints that the vehicle travels through cannot obtain the corresponding positions in the actual road network, resulting in a partial loss of the travel trajectory. This part of the checkpoints is referred to as invalid checkpoints in this article. In order to solve the problem of incomplete trajectories caused by invalid checkpoints, in addition to finding and deleting these invalid checkpoints in a single travel trajectory, considering the connectivity of the trajectory, the trajectory restoration process should also be performed to finally obtain a complete vehicle travel trajectory. Furthermore, in the step S5, the following process is included:

[0092] According to the row and column labels of the adjacency matrix of the checkpoint network topology structure and the geographical location information of the checkpoint device, the longitude and latitude data of the checkpoint vertex corresponding to the checkpoint network topology structure in the actual road network are queried; if the corresponding longitude and latitude data cannot be queried, the checkpoint identified by the checkpoint number is identified as an invalid checkpoint; the invalid checkpoints and their corresponding association function relationships in the checkpoint network topology structure and its adjacency matrix are deleted;

[0093] Traverse the single trip trajectory SR in order to find the invalid checkpoints in the single trip trajectory. According to the position of the invalid checkpoints in the travel trajectory, they are divided into the following two cases for processing:

[0094] If the invalid checkpoint is a single trip trajectory SR = {b1, b2, ..., b i-1 ,bi ,b i+1 ,…,b D} starting point b1 or end point b D , then directly delete the invalid card port;

[0095] If the invalid checkpoint is the travel trajectory SR={b1,b2,…,b i-1 ,b i ,b i+1 ,…,b D}Intermediate bayonet b i , use Dijkstra algorithm to find the card port b on the card port network topology structure i-1 and b i+1 The shortest path between If the shortest path exists, the invalid checkpoints in the travel trajectory are replaced with the intermediate nodes of the shortest path in sequence; otherwise, the invalid checkpoints are directly deleted.

[0096] More specifically, the latitude and longitude coordinates of the invalid checkpoints are derived from the traffic checkpoint equipment distribution map. The OSM city road network map is required to use the Dijkstra algorithm to find the shortest path between two checkpoints.

[0097] The following will provide further explanation based on specific data.

[0098] The urban road network data and traffic checkpoint equipment distribution of this embodiment are as follows: Figure 4 As shown, the total mileage of roads in Foshan is 7018 kilometers. The data source in the embodiment is the local diesel truck checkpoint data from May 24 to May 30, 2021 in Foshan, totaling 1,621,544 original vehicle passing records from 638 traffic checkpoint devices.

[0099] In step S1, the attributes such as the checkpoint number, the vehicle passing date and the vehicle passing time in the original checkpoint data are first formatted and modified, as shown in Table 1; then the checkpoint passing records without identifying the vehicle license plate or checkpoint number and the repeated vehicle passing records are screened. The processed checkpoint data contains 1,616,086 vehicle passing records.

[0100] Table 1 Data format conversion example

[0101] Attribute columns name Original data form Standardized data format KKBH Mount number 440604305050002000 "440604305050002000” GQ Date of passing 2021-05-21 07 2021-05-21 GCSJ Passing time 2021-05-21 07:56:59.0000000 2021-05-21 07:56:59

[0102] In the step S2, the bayonet data is divided into groups with the vehicle license plate as the unique identifier, and sorted in ascending order according to the passing time of the vehicle, forming a set of single-vehicle travel information. In this embodiment, taking a vehicle with the license plate "粤E****b" and the passing date "2021-05-25" as an example, the trajectory reconstruction operation is carried out. The set of single-vehicle travel information of this vehicle contains 9 bayonet passing records, as shown in Table 2.

[0103] Table 2 Bayonet Data of Vehicle "粤E****b"

[0104]

[0105]

[0106] In the step S3, the bayonet network topology structure of Foshan City is constructed:

[0107] Adjacent bayonet pair identification: Extract the bayonet pairs with adjacent passing times in all sets of single-vehicle travel information and retain the high-frequency bayonet pairs as adjacent bayonet pairs, a total of 2988 pairs; then calculate the driving time of the 55th percentile vehicle on the road section as the average travel time at of the adjacent bayonet pairs. i,j 。

[0108] Adjacency matrix construction: According to the actual operation status of the road network in Foshan City and daily life experience, the minimum red light time R is taken as 15s, and the travel time threshold B is taken as the 85th percentile road network travel time, that is, 6380s. Then, filter the adjacent bayonet pairs where at i,j ≤R and at i,j > and mark them. The marked abnormal bayonet pairs account for 15.03% of all bayonet pairs. Finally, according to the marked abnormal bayonet pairs, construct the bayonet network topology structure and its adjacency matrix, as Figure 5 shown.

[0109] In the step S3, since the driving time of the vehicle between the bayonets numbered "440615000000212021" and "440615000000126015" exceeds the travel time threshold B, this trajectory is divided into two single trips, that is:

[0110] Trip 1: 440615000000224001-440615000000226068-440615000000226067-440615000000312004-440615000000226015-440615000000212021;

[0111] Trip 2:440615000000126015-440615000000412004-440615000000124001.

[0112] In step S4, the single travel trajectory of the vehicle is matched with the address and restored:

[0113] Address matching and invalid checkpoint identification: Match the vehicle's single trip trajectory with the actual road network map, and there are 3 invalid checkpoints. The travel trajectory between location numbers 3-7 in trip 1 and location numbers 7-1 in trip 2 is incomplete, such as Figure 6 As shown in the figure, delete the three invalid bayonet vertices and their associated functions in the bayonet network topology.

[0114] The Dijkstra algorithm is used to restore the travel trajectory after deleting the invalid checkpoints: After deleting the invalid checkpoints from the vehicle's single travel trajectory, based on the checkpoint network topology, the checkpoints at location numbers 3 and 7 are used as the starting and ending points in trip 1, and the checkpoints at location numbers 7 and 1 are used as the starting and ending points in trip 2. The Dijkstra algorithm is used to find the shortest travel time path, and the travel trajectory after deleting the invalid checkpoints is restored. The reconstructed trajectory is as follows Figure 7 shown.

[0115] 1797 vehicle travel trajectories with GPS data were selected from the reconstructed trajectories for verification, and trajectory buffers of different radii were constructed. Then, the coincidence rate with the GPS data points was calculated to measure the similarity between the two. When the buffer radius was 0.001, 0.005, 0.01, and 0.02 decimal degrees, the calculation results were as follows: Figure 8 When the radius is greater than 0.02, the coincidence rate between most of the verified trajectories and the GPS data exceeds 80%, indicating that the trajectory reconstruction result is close to the actual GPS data, thus verifying the effectiveness and feasibility of the method in this embodiment.

[0116] Example 2

[0117] A computer storage medium stores a computer program, which, when executed by a processor, implements the steps of the vehicle travel trajectory reconstruction method considering the checkpoint network topology structure in embodiment 1.

[0118] Example 3

[0119] A computer device includes a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor. When the computer program is executed by the processor, the steps of the vehicle travel trajectory reconstruction method considering the checkpoint network topology structure in Example 1 are implemented.

[0120] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A vehicle travel trajectory reconstruction method considering the topological structure of the checkpoint network, characterized in that: The following steps are involved: S1, obtaining original card slot data, adjusting the format of the original card slot data, and filtering out invalid data; S2, generating a bicycle travel information set with a time-related order according to the processing result of step S1; S3, constructing a checkpoint network topology structure based on the bicycle travel information set and the actual traffic rules of the road network nodes; the checkpoint network topology structure is a weighted directed graph describing the spatiotemporal relationship between the road network checkpoints, and is used to provide a geographic reference for subsequent address matching and trajectory restoration operations; In step S3, an adjacency matrix is ​​constructed in the following manner to mark the extracted adjacent checkpoint pairs as outliers: Extract all the ports contained in the adjacent port pair As the row and column labels of the adjacency matrix, m is the number of checkpoints contained in the list of adjacent checkpoint pairs, then the size of the adjacency matrix is m × m ; According to the preset red light time threshold R， , and select 70%~90% of the road network travel time as the travel time threshold B ;Will and The adjacent bayonet pair Mark as abnormal bayonet pair; Determine the weight of the adjacency matrix If the value of , then it means the bayonet , Not directly connected, so ;like ,but ; S4, dividing a single trip trajectory according to the bicycle travel information set and a preset travel time threshold; S5, performing address matching and trajectory restoration according to the checkpoint network topology structure and a single travel trajectory.

2. The vehicle travel trajectory reconstruction method considering the checkpoint network topology structure according to claim 1 is characterized in that: In step S1: Obtain OSM city road network maps and traffic checkpoint equipment distribution maps, and collect original checkpoint data detected by road network checkpoint equipment within a specific time period; According to the preset format standard, the attributes of the original checkpoint data including the checkpoint number, the date of vehicle passing and the time of vehicle passing are formatted; the checkpoint passing records in which the vehicle license plate or the checkpoint number is not identified in the original checkpoint data, as well as the duplicate vehicle passing records are filtered out.

3. The vehicle travel trajectory reconstruction method considering the checkpoint network topology structure according to claim 1 is characterized in that: In step S2, a week is used as a statistical time node to determine the single-vehicle travel information of a vehicle within a week: The processing result of step S1 is divided into groups identified by vehicle license plates, each group including all vehicle passing records at different times; Sort each group in ascending order according to the detected passing time to generate a set of bicycle travel information with time-related order ;in, N is the number of times a vehicle is detected in the checkpoint data within a week, This is a video clip of a vehicle passing through a checkpoint. is the mount number corresponding to the mount device. The date when the vehicle passes by the camera device. is the corresponding passing time, and .

4. The vehicle travel trajectory reconstruction method considering the checkpoint network topology structure according to claim 3 is characterized in that: In step S3, The process includes: According to the bicycle travel information set and in combination with the actual traffic rules of the road network nodes, high-frequency checkpoint pairs that are adjacent in time are extracted as adjacent checkpoint pairs, and the average travel time of vehicles passing through the adjacent checkpoint pairs is calculated; An adjacency matrix is ​​constructed according to the adjacent bayonet pairs.

5. The vehicle travel trajectory reconstruction method considering the checkpoint network topology structure according to claim 4 is characterized in that: In step S3, high-frequency checkpoint pairs that are close in time are extracted as adjacent checkpoint pairs in the following manner, and the average travel time of vehicles passing through the adjacent checkpoint pairs is calculated: According to the bicycle travel information collection The time of each vehicle passing record in the system is used to identify all the time-related checkpoints. , that is, the bayonet numbers of adjacent bayonet ports in the set are connected in sequence; With the preset minimum support SupSUM As the threshold, the statistical card pair Frequency ;like , then keep the bayonet pair as the adjacent bayonet pair; Calculate the travel time of all vehicles passing through adjacent checkpoint pairs ; The 55th percentile travel time on the road section is selected as the adjacent checkpoint pair. , Average travel time between ,Right now ;in, For vehicles k Passing the bayonet The time when the vehicle passes is detected. For vehicles k Passing the bayonet The time when the vehicle passes is detected. To pass through the adjacent bayonet The travel time set of all vehicles is , n It is the number of vehicles passing through the checkpoint pair within a specific time period.

6. The vehicle travel trajectory reconstruction method considering the checkpoint network topology structure according to claim 1 is characterized in that: In step S4, the single trip trajectory is divided by repeating the following steps: Traverse the bicycle travel information collection in order ; Compare vehicles passing through checkpoints The date of the vehicle passing , Whether they are consistent: If the dates are different and the vehicle travels through the checkpoint pair , it means that the vehicle is at the checkpoint There is a long period of static state between them, that is, the travel trajectories before and after this checkpoint pair do not belong to the same single trip; According to the adjacent bayonet Average travel time between And the marked abnormal checkpoints are used to determine the vehicle operation status: if , it means that the checkpoint pair does not match the actual distance that the vehicle can move at normal speed or is a duplicate record caused by the waiting time for the red light at the intersection, so only the adjacent checkpoint pairs are retained. The next checkpoint passing record in the ;like , indicating that the vehicle is stationary for a long time between the adjacent checkpoints; The adjacent checkpoints where the vehicles are stationary for a long time in the bicycle travel information set are Split to generate two consecutive single trips, namely , .

7. The vehicle travel trajectory reconstruction method considering the checkpoint network topology structure according to claim 6 is characterized in that: In step S5, The process includes: According to the row and column labels of the adjacency matrix of the checkpoint network topology structure and the geographical location information of the checkpoint device, the longitude and latitude data of the checkpoint vertex corresponding to the checkpoint network topology structure in the actual road network are queried; if the corresponding longitude and latitude data cannot be queried, the checkpoint identified by the checkpoint number is identified as an invalid checkpoint; the invalid checkpoints and their corresponding association function relationships in the checkpoint network topology structure and its adjacency matrix are deleted; Traverse a single trip trajectory in order SR , find the invalid checkpoints in a single travel trajectory, and handle them in the following two cases according to the location of the invalid checkpoints in the travel trajectory: If the invalid checkpoint is a single trip trajectory The starting point Or the terminal , then directly delete the invalid card port; If the invalid checkpoint is a travel trajectory Middle bayonet , use Dijkstra algorithm to find the checkpoints on the checkpoint network topology and The shortest path between , if the shortest path exists, the invalid checkpoints in the travel trajectory are replaced with the intermediate nodes of the shortest path in sequence; otherwise, the invalid checkpoints are directly deleted.

8. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the vehicle travel trajectory reconstruction method considering the checkpoint network topology structure as described in any one of claims 1 to 7 are implemented.

9. A computer device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor, characterized in that: When the computer program is executed by a processor, the steps of the vehicle travel trajectory reconstruction method considering the checkpoint network topology structure as described in any one of claims 1 to 7 are implemented.

Images