Traffic jam sensing method and system based on ETC portal frame data

Abstract

The invention relates to the technical field of intelligent traffic, and provides a traffic jam sensing method and system based on ETC portal data, and the method comprises the steps: dividing a road segment based on an ETC portal, and determining a bidirectional portal sequence; identifying daily high-frequency portal pairs according to the vehicle passing amount, screening the daily high-frequency portal pairs, constructing a road network congestion fusion sensing portal sequence library, and updating the road network congestion fusion sensing portal sequence library; activating a daily high-frequency portal pair of the road network congestion fusion sensing portal sequence library, and based on road section passing time, performing combined sensing on the activated daily high-frequency portal pair by adopting a slope analysis method and a standard deviation analysis method to determine a jammed portal pair road section; and aiming at the road section which is determined to be congested, carrying out image identification through video monitoring data, and determining a specific congested road section. According to the invention, the approximate congestion road section is identified through the ETC door frame data, the pile number of the congestion road section is accurately identified through the video monitoring data, and the congestion event can be quickly found and handled.

Description

Technical Field

[0001] The present disclosure relates to the technical field related to intelligent transportation, and in particular, to a traffic congestion perception method and system based on ETC gantry data. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] With the increase in traffic volume on highways, especially in special weather conditions or when traffic accidents occur, congestion often occurs in the main sections and toll station ramps of highways, seriously affecting the efficiency and safety of highway traffic. By sensing abnormal road conditions in real time and shortening the time cycle of "discovery-response-disposal", precise congestion relief can be achieved, effectively reducing the duration and scope of congestion, and ensuring smooth operation.

[0004] Highway surveillance video recognition is one of the important ways to automatically perceive congestion in real time. Based on computer vision and deep learning technology, it analyzes features such as vehicle operation and road vehicle density in video images to determine whether abnormal road conditions occur. However, due to factors such as multi-angle monitoring of highways and variable natural lighting conditions, video recognition algorithms have some limitations, and there are a certain number of false alarms in practice; in addition, due to limited video coverage, real-time congestion perception cannot be achieved on some sections of the road. At present, manual reporting of congestion still uses common means, generally through patrol personnel or car owners to report events, and then verified by operation managers and reported to the system, but this method has strong subjective factors and time lags. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a traffic congestion perception method and system based on ETC gantry data, which utilizes highway ETC gantry data to explore the vehicle operation characteristics of congested road sections and perceive congested road conditions in real time and comprehensively.

[0006] In order to achieve the above objectives, the present disclosure adopts the following technical solutions:

[0007] One or more embodiments provide a traffic congestion perception method based on ETC gantry data, comprising the following steps:

[0008] Divide the road sections based on the ETC gantries, match the upstream and downstream gantries in each section, and determine the bidirectional gantries sequence according to the flow order;

[0009] Identify the daily high-frequency gantry pairs according to the vehicle volume, select the daily high-frequency gantry pairs in all the bidirectional gantry sequences of the road network, build a road network congestion fusion perception gantry sequence library, and remove the invalid gantry pairs in real time for updating;

[0010] Activate the daily high-frequency gantry pairs in the road network congestion fusion perception gantry sequence library, use slope analysis and standard deviation analysis to perform joint perception of general congestion based on the road section travel time, use zero-pass judgment to perceive severe congestion, and determine the congested gantry pair sections;

[0011] To identify the congested gantry sections, image recognition is performed through video surveillance data to determine the specific congested sections.

[0012] One or more embodiments provide a traffic congestion perception system based on ETC gantry data, including:

[0013] Road section division module: It is configured to divide the road sections based on the ETC gantries, match the upstream and downstream gantries in each road section, and determine the bidirectional gantries sequence according to the flow order;

[0014] Gantry sequence library construction module: It is configured to identify the daily high-frequency gantry pairs according to the vehicle volume, filter the daily high-frequency gantry pairs in all the bidirectional gantry sequences of the road network, build a road network congestion fusion perception gantry sequence library, and remove the invalid gantry pairs in real time for updating;

[0015] Congestion section identification module: It is configured to activate the daily high-frequency gantry pairs in the road network congestion fusion perception gantry sequence library according to the set activation conditions, and use the slope analysis method and standard deviation analysis method to perform general congestion joint perception on the activated daily high-frequency gantry pairs based on the road section travel time, and use zero-passing vehicle judgment to perceive severe congestion and determine the congested gantry pair sections;

[0016] Congestion location precise positioning module: It is configured to determine the congested gantry section, perform image recognition through video surveillance data, and determine the specific congested section.

[0017] An electronic device includes a memory and a processor, and computer instructions stored in the memory and running on the processor. When the computer instructions are run by the processor, the steps in the above-mentioned traffic congestion perception method based on ETC gantry data are completed.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] The present disclosure uses ETC gantry data to identify the approximate congested sections, explores the distribution characteristics of abnormal vehicle travel time in the sections between gantry pairs, perceives abnormal highway conditions in real time and comprehensively, and further verifies the perception results and focuses on congested sections in combination with video surveillance data. It can accurately identify the pile numbers of congested sections through video surveillance data, and can achieve rapid discovery and rapid disposal of congestion events. It can effectively identify traffic congested sections and accurately locate the specific location of congestion through video surveillance data, thereby improving the accuracy and real-time perception of traffic congestion. By dynamically updating high-frequency gantry pairs, the timeliness and effectiveness of data are guaranteed, and the efficiency of traffic management is improved.

[0020] The advantages of the present disclosure and the advantages of additional aspects will be described in detail in the following specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The exemplary embodiments of the present disclosure and the description thereof are used to explain the present disclosure but do not constitute a limitation of the present disclosure.

[0022] Figure 1 is a flow chart of a traffic congestion perception method based on ETC gantry data according to Embodiment 1 of the present disclosure; DETAILED DESCRIPTION

[0023] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0024] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0025] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof. It should be noted that, in the absence of conflict, the various embodiments in the present disclosure and the features in the embodiments can be combined with each other. The embodiments will be described in detail below in conjunction with the accompanying drawings.

[0026] Example 1

[0027] In the technical solutions disclosed in one or more embodiments, Figure 1 As shown, a traffic congestion perception method based on ETC gantry data includes the following steps:

[0028] Step 1: Divide the road sections based on the ETC gantries, match the upstream and downstream gantries in each section, and determine the bidirectional gantries sequence according to the flow order;

[0029] Step 2: Identify the daily high-frequency gantry pairs according to the vehicle volume, select the daily high-frequency gantry pairs in all the bidirectional gantry sequences of the road network, build a road network congestion fusion perception gantry sequence library, and remove the invalid gantry pairs in real time for updating;

[0030] Step 3: Activate the daily high-frequency gantry pairs in the road network congestion fusion perception gantry sequence library according to the set activation conditions. Based on the road section travel time, use the slope analysis method and the standard deviation analysis method to perform general congestion joint perception on the activated daily high-frequency gantry pairs. Use zero-pass judgment to perceive severe congestion and determine the congested gantry pair sections.

[0031] Step 4: For the congested gantry sections, image recognition is performed based on deep learning through video surveillance data to determine the specific congested sections.

[0032] The ETC gantry on the highway can record the vehicle's driving path for use in calculating highway tolls. In addition, the massive amount of structured data it records, such as vehicle identification codes, vehicle types, driving speeds, passing times, toll stations entering the highway, etc., can be easily integrated with video surveillance, millimeter-wave radar and other data to achieve in-depth application of ETC gantry data.

[0033] In this embodiment, the approximate congested sections are identified through ETC gantry data, and the abnormal vehicle travel time distribution characteristics of the sections between gantry pairs are excavated to perceive the abnormal road conditions of the highway in real time and comprehensively. The perception results are further verified and the congested sections are focused on in combination with video surveillance data. The pile numbers of the congested sections are accurately identified through video surveillance data, which can achieve rapid discovery and rapid disposal of congestion events. It can effectively identify traffic congested sections and accurately locate the specific location of congestion through video surveillance data, thereby improving the accuracy and real-time perception of traffic congestion. By dynamically updating high-frequency gantry pairs, the timeliness and effectiveness of data are guaranteed, and the efficiency of traffic management is improved.

[0034] An alternative technical solution, the method can also adjust the activation conditions and analysis methods according to different traffic flows and road conditions. For example, different slope and standard deviation thresholds can be used to adapt to different traffic conditions. In addition, other types of traffic data, such as GPS data and traffic sensor data, can be introduced to further improve the accuracy of traffic congestion perception.

[0035] In step 1, the road section is divided based on the ETC gantry;

[0036] Toll stations, hub interchanges and other locations are usually key nodes for changes in traffic volume. It is necessary to appropriately interrupt the road sections at the above-mentioned locations and match the ETC gantries to the road sections in order to more accurately reflect the topological relationship of the road network and the degree of adaptability of the gantry data to the road network.

[0037] Specifically, the method for dividing the road sections based on the ETC gantry and matching to obtain the bidirectional gantry sequence includes the following steps:

[0038] Step 11: Based on the vehicle passing information collected by the ETC gantry, statistical analysis is performed on the location, flow direction, traffic volume and other data of each toll station and hub interchange to identify the traffic flow distribution and main flow direction of the road network and determine the key nodes and sections where traffic volume changes;

[0039] Step 12: Disconnect the road network at key nodes to obtain divided road sections;

[0040] Specifically, identify key nodes and sections, follow topological principles, consider the actual highway construction situation, interrupt the road network and sections at key nodes, ensure that there will be no flow changes in the interrupted sections, and ensure the logic and consistency of the graphics;

[0041] Step 13: For the divided road sections, establish the connection relationship between the road sections according to the topological principle, construct a road network topology map, match the ETC gantries of the corresponding road sections to the road sections, and form a bidirectional gantries sequence according to the order of vehicle flow;

[0042] Specifically, the ETC gantry's name, longitude and latitude and other information are matched with the road sections in the road network, and the ETC gantry is matched to the corresponding road section according to the position, so as to obtain the road section division and upstream and downstream gantries in the application network, and form a bidirectional gantry sequence in the order of flow direction.

[0043] The solution in step 1 above can accurately divide the road sections and match the ETC gantries to form a two-way gantries sequence, providing basic data for subsequent traffic congestion perception. By identifying key nodes and traffic distribution, traffic flow can be monitored and managed more effectively.

[0044] In step 2, a road network congestion fusion perception gantry sequence library is constructed;

[0045] The application network of congestion fusion perception is usually large in scale and has a large number of bidirectional gantries. For example, in the provincial highway network, it takes a lot of time to calculate one by one. At the same time, the ETC gantries sometimes record the opposite vehicles by mistake, affecting the quality of the vehicle passing records of the gantry pair. In addition, when the ETC gantry monitoring equipment is damaged due to age and cannot record, the gantry pair will be unavailable. Therefore, in order to ensure the timeliness, comprehensiveness and accuracy of perception, it is necessary to properly screen the gantries of the entire road network.

[0046] In addition, for each gantry pair, the normal travel time calculated based on historical vehicle passing data is not fixed due to the influence of seasons, social activities, holidays, etc. The gantry pairs in the road network congestion fusion perception gantry sequence library need to be dynamically adjusted according to real-time conditions.

[0047] In step 2, the method for constructing and updating the road network congestion fusion perception gantry sequence library includes the following steps:

[0048] Step 21, counting the number of vehicles passing through the gantry pair every day;

[0049] Specifically, all bidirectional gantry sequences in the application network obtained in step 1 are traversed every day, and the number of vehicles passing through all bidirectional gantry sequences in the network on the previous day is counted to obtain the number of vehicles passing through each day;

[0050] Step 22: According to the obtained daily vehicle passing number, the daily high-frequency gantry pairs are screened out according to the set conditions; specifically, the steps include:

[0051] Step 221, sort all bidirectional gantry sequences in ascending order according to the number of vehicles passing through each day, and dynamically extract gantry pairs whose number of vehicles passing through is not less than the lower quartile of the number of vehicles passing through all gantry sequences each day;

[0052] Step 222: for the extracted gantry pairs, the gantry pairs with abnormal flow directions are eliminated by using the matching relationship between the gantry pairs and the road sections in the road network and the topological relationship, and the remaining gantry pairs are used as the high-frequency gantry pairs of the day;

[0053] Step 223, dynamically remove the gantry pairs whose rankings have been lower than the lower quartile of the number of vehicles passing through the entire gantry sequence on the same day for three consecutive days among the high-frequency gantry pairs of the day, and obtain the high-frequency gantry pairs of the corresponding day;

[0054] In the above step 22, it is ensured that the selection of high-frequency gantry pairs is based on the actual vehicle passing data, eliminating the interference of abnormal flow direction, and improving the accuracy and reliability of congestion perception. Low-frequency gantry pairs are dynamically eliminated to maintain the timeliness of the data.

[0055] Alternative implementation methods may adjust the quartile threshold according to different traffic conditions and needs, or introduce other statistical analysis methods, such as median and standard deviation, to improve the flexibility and accuracy of data screening.

[0056] Step 23: according to the actual number of vehicles passing through multiple time windows of the high-frequency gantry pair, activate the high-frequency gantry pair and build a road network congestion fusion perception gantry sequence library;

[0057] Specifically, as time goes by, it is determined whether the number of vehicles passing through each high-frequency gantry pair in the current time window is not less than the vehicle passing threshold c. When the number of time windows that meet the above conditions is equal to 15 time windows, the gantry pair is activated, and the travel time of all vehicles in the 15 time windows is obtained. In order to avoid the impact of vehicles entering service areas or abnormal parking on perception, the abnormal travel time greater than the upper quartile is eliminated.

[0058] Among them, the threshold value of the number of vehicles passing, c, is taken as the lower quartile of the average number of vehicles passing in the time window of the previous day. For special periods such as early morning and bad weather, this value can be appropriately reduced; the time window size can be set;

[0059] Step 24, construct a first-in-first-out dynamic queue, store the data of the time window of the activated gantry pair, and update it in real time according to the update data of the high-frequency gantry pair;

[0060] The data of the time window of the activated gantry pair include: the time points of all vehicles passing through the set number of time windows, the passing time T, and the average value t of the passing time of all vehicles in the set number of time windows;

[0061] Optionally, a first-in-first-out dynamic queue is maintained to store the 15 time windows in step 23, and a normal travel time set T is maintained to store the travel time of all vehicles in the 15 time windows in step 23, and the average vehicle travel time t is calculated for use in congestion fusion perception, and the vehicle passing time point, set T and average t are continuously updated over time;

[0062] Step 25: Calculate the travel time of activated high-frequency gantry pairs, identify high-frequency gantry pairs with abnormal travel time changes, and remove them from the road network congestion fusion perception gantry sequence library;

[0063] Optionally, a method for identifying high-frequency gantry pairs with abnormally changing travel time is to identify the gantry pairs whose earliest and latest timestamps of the travel time set T exceed a set time, such as 3 hours.

[0064] Specifically, if among the activated gantry pairs, there is a gantry pair set T whose earliest and latest timestamp time interval exceeds 3 hours, then the activated gantry pair meets the expiration condition and is deleted from the activated gantry pairs.

[0065] The method in step 2 above can timely identify and update high-frequency gantry pairs, ensuring the accuracy and real-time perception of road network congestion. The efficiency and accuracy of data processing are improved through the construction and real-time update of dynamic queues.

[0066] Alternatively, different data storage and update strategies can be used, such as distributed storage and processing based on cloud computing, to improve the efficiency and scalability of data processing. In addition, machine learning algorithms can be introduced to further optimize the identification and update strategies of high-frequency gantry pairs.

[0067] In step 3, the congestion fusion perception method based on the road section travel time includes a general congestion identification method and a severe congestion identification method set in parallel, which are specifically described as follows;

[0068] By extracting the vehicle passing data of the upstream and downstream gantries of the road section, the vehicle travel time is calculated, and the abnormal changes in the travel time are used to determine whether the road section is congested. According to the degree of congestion, it can be divided into general congestion and severe congestion events. The two congestion situations are judged simultaneously, and congestion can be reported at the earliest moment of perception.

[0069] Optionally, a general congestion identification method includes the following steps:

[0070] Step 31, traverse each activated high-frequency gantry pair, obtain the travel time of all vehicles in the current time window and the previous set number of time windows, and store them in the current travel time set T′;

[0071] In this embodiment, the number of time windows set may be 15, that is, the travel time of all vehicles in the current time window and the previous 15 time windows is obtained to obtain the current travel time set T′;

[0072] Step 32: for the travel time of all vehicles in each window of the activated high-frequency gantry, the slope analysis method and the standard deviation analysis method are respectively used to perceive and identify the approximate congested road section, that is, the congested gantry pair road section;

[0073] Generally, congestion events will cause the travel time of most vehicles between gantries to increase. For such events, the slope analysis method and the standard deviation analysis method are used for joint perception.

[0074] In some embodiments, a slope analysis method is used, with time as the horizontal axis and travel time as the vertical axis, to determine whether a general congestion event occurs based on the slope change;

[0075] In some embodiments, for each activated gantry pair, a slope analysis method is used to perform the following steps:

[0076] Step A1: Divide the travel time in the current travel time set T′ by the corresponding mean of the travel time to obtain a first multiple set B′;

[0077] Divide the travel time in the normal travel time set T by the corresponding mean of the travel time to obtain a second multiple set B;

[0078] The above method can achieve dimensional unification, and the travel time in the current travel time set T′ is divided by the mean travel time of the gantry pair in the time period. Get the multiple set B′, divide the travel time in the normal travel time set T by the mean travel time of the gantry pair in the time period Get multiple set B;

[0079] Step A2: Calculate the first slope of the first multiple set B′ with respect to time, denoted as K1; calculate the second slope of the second multiple set B with respect to time, denoted as K2

[0080] That is: calculate the slope of the current time window and the previous 15 time window multiples set B′ to time, recorded as K1, calculate the slope of the normal travel time multiples set B to time, recorded as K2;

[0081] Step A3: When the difference between the first slope and the second slope exceeds the set value, it is judged as general congestion, and the standard deviation analysis method is performed to make the next judgment; otherwise, it is judged as no congestion, and the process ends;

[0082] Specifically, if K1-K2>K j , then the next step can be entered, otherwise, the section between the gantry pairs is not generally congested;

[0083] Among them, K j is the difference in slope, and the reference value is 0.8;

[0084] In order to reduce the impact of abnormally moving vehicles on perception, the time with more concentrated travel time distribution is extracted, and the standard deviation analysis method is used to further improve the perception accuracy.

[0085] In this embodiment, the gantry pair mean refers to the ETC gantry pair;

[0086] In some embodiments, for the gantry pairs that are confirmed to be generally congested by the slope analysis method, the standard deviation analysis method is used to further determine the congestion situation, and the specific steps are as follows:

[0087] Step B1, setting the standard deviation threshold of the passage time of the gantry pair in each time window, denoted as σ0, σ0=0.2μ0, μ0 is the mean of the passage time of the gantry sequence in the time window;

[0088] Step B2, calculating the standard deviation of the passage time of the gantry pair in each time window, denoted as σ;

[0089] Step B3: If σ≤σ0, it can be determined that the road section between the gantry pair is generally congested; otherwise, the current gantry pair is not congested;

[0090] In the above scheme, the combined use of slope analysis and standard deviation analysis can more accurately judge the congestion situation, reduce misjudgment, and improve the accuracy and reliability of congestion perception.

[0091] Severe congestion will lead to serious traffic jams. Vehicles stop on the highway, so no vehicles pass through the downstream gantry for a certain period of time. Zero vehicle passing is used to determine whether it is severe congestion. The following specific steps are performed for the activated gantry:

[0092] Step C1: for the current gantry pair, if the number of vehicles passing in the two time windows before the current time window is ≥ c, proceed to the next step C2 for judgment; otherwise, the road section between the current gantry pair is not severely congested;

[0093] Step C2: For the current gantry pair, if the time difference between the last vehicle passing time in the previous time window and the end time window is greater than or equal to the set time length, it is judged to be severely congested; otherwise, the road section between the current gantry pair is not severely congested;

[0094] Optionally, the set time length can be 3 minutes. If the last vehicle passing time in the previous time window is not less than 3 minutes from the end of the time window, then if no vehicle passes the downstream gantry in the section between the gantry pairs for more than 3 minutes, it can be judged as severe congestion. Otherwise, the section between the gantry pairs is not severely congested.

[0095] In step 4, accurate identification of congested road sections based on the fusion of ETC gantry data and video surveillance data is achieved, which specifically includes the following steps:

[0096] Step 41, according to the identified congested gantry road section, determine the road section according to the road network topology map;

[0097] Step 42, obtaining video data of the road section corresponding to the congested gantry;

[0098] Step 43, performing frame processing on the video image to obtain a frame image;

[0099] Step 44: Perform image recognition on the acquired frame image based on the deep learning model to determine the pile number of the congested road section.

[0100] Optionally, deep learning models such as YOLO model, Faster R-CNN, or DeepLab can be used based on deep learning models;

[0101] Specifically, determining the pile number of the congested road section includes the following steps:

[0102] Step 441: Use the trained deep learning model to identify the preprocessed frame images and determine the congested area in each frame image;

[0103] Step 442: Identify areas with obvious congestion based on characteristics such as vehicle density and vehicle speed;

[0104] Step 443: based on the identified congested area, convert the geographical location information corresponding to the video frame image into a road section stake number;

[0105] Step 444: Perform aggregation processing based on the results of multiple frames of images to determine the start and end pile numbers of the congested road section.

[0106] Since the average spacing between ETC gantry intervals is about 8 kilometers and the longest length can exceed 15 kilometers, in this embodiment, for the ETC gantry sections perceived as congested in step 3, the congestion image recognition through video surveillance is requested to further determine the pile number of the congested section, thereby achieving congestion perception focusing.

[0107] Example 2

[0108] Based on Example 1, this embodiment provides a traffic congestion perception system based on ETC gantry data, including:

[0109] Road section division module: It is configured to divide the road sections based on the ETC gantries, match the upstream and downstream gantries in each road section, and determine the bidirectional gantries sequence according to the flow order;

[0110] Gantry sequence library construction module: It is configured to identify the daily high-frequency gantry pairs according to the vehicle volume, filter the daily high-frequency gantry pairs in all the bidirectional gantry sequences of the road network, build a road network congestion fusion perception gantry sequence library, and remove the invalid gantry pairs in real time for updating;

[0111] Congestion section identification module: It is configured to activate the daily high-frequency gantry pairs in the road network congestion fusion perception gantry sequence library according to the set activation conditions, and use the slope analysis method and standard deviation analysis method to perform general congestion joint perception on the activated daily high-frequency gantry pairs based on the road section travel time, and use zero-passing vehicle judgment to perceive severe congestion and determine the congested gantry pair sections;

[0112] Congestion location precise positioning module: It is configured to determine the congested gantry section, perform image recognition through video surveillance data, and determine the specific congested section.

[0113] It should be noted here that the various modules in this embodiment correspond one-to-one to the various steps in Example 1, and the specific implementation process is the same, which will not be repeated here.

[0114] Example 3

[0115] Based on Example 1, this embodiment provides an electronic device, including a memory and a processor, and computer instructions stored in the memory and running on the processor. When the computer instructions are executed by the processor, the steps in the traffic congestion perception method based on ETC gantry data described in Example 1 are completed.

[0116] The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, the present disclosure may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

[0117] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A traffic congestion perception method based on ETC gantry data, characterized in that: The steps include: Divide the road sections based on the ETC gantries, match the upstream and downstream gantries in each section, and determine the bidirectional gantries sequence according to the flow order; Identify the daily high-frequency gantry pairs according to the vehicle volume, select the daily high-frequency gantry pairs in all the bidirectional gantry sequences of the road network, build a road network congestion fusion perception gantry sequence library, and remove the invalid gantry pairs in real time for updating; Activate the daily high-frequency gantry pairs in the road network congestion fusion perception gantry sequence library, use slope analysis and standard deviation analysis to perform joint perception of general congestion based on the road section travel time, use zero-pass judgment to perceive severe congestion, and determine the congested gantry pair sections; To identify the congested gantry sections, image recognition is performed through video surveillance data to determine the specific congested sections.

2. The traffic congestion perception method based on ETC gantry data according to claim 1, characterized in that: The method for dividing the road sections based on the ETC gantry and matching to obtain a bidirectional gantry sequence includes the following steps: Based on the vehicle passing information collected by the ETC gantries, the location, flow direction, and traffic volume data of each toll station and hub interchange are statistically analyzed to identify the traffic flow distribution and main flow direction of the road network and determine the key nodes and sections where traffic volume changes; The road network is interrupted at key nodes to obtain divided road sections; For the divided road sections, the connection relationship between the sections is established according to the topological principles, and a road network topology map is constructed. The ETC gantries of the corresponding sections are matched to the sections, and a bidirectional gantry sequence is formed according to the order of vehicle flow.

3. The traffic congestion perception method based on ETC gantry data according to claim 1, characterized in that: The method for constructing and updating a road network congestion fusion perception gantry sequence library includes the following steps: Count the number of vehicles passing through the gantry pair every day; According to the obtained daily vehicle passing number, filter out the daily high-frequency gantry pairs according to the set conditions; According to the actual number of vehicles passing through multiple time windows of the high-frequency gantry pairs, the activated high-frequency gantry pairs are used to build a road network congestion fusion perception gantry sequence library; Construct a first-in-first-out dynamic queue to store the data of the time window of the activated gantry pairs and update it in real time according to the update data of the high-frequency gantry pairs; Calculate the travel time of activated high-frequency gantry pairs, identify high-frequency gantry pairs with abnormal travel time changes, and remove them from the road network congestion fusion perception gantry sequence library.

4. The traffic congestion perception method based on ETC gantry data as claimed in claim 3, characterized in that: According to the set conditions, filter out the daily high-frequency gantry pairs. The steps are as follows: Sort all bidirectional gantry sequences in ascending order according to the number of vehicles passing through each day, and dynamically extract gantry pairs whose number of vehicles passing through is not less than the lower quartile of the number of vehicles passing through all gantry sequences every day; For the extracted gantry pairs, the matching relationship between the gantry pairs and the road sections in the road network is used, and the gantry pairs with abnormal flow directions are eliminated using the topological relationship, and the remaining gantry pairs are used as the high-frequency gantry pairs of the day; Dynamically eliminate the gantry pairs whose ranking among the high-frequency gantry pairs of the day has been lower than the lower quartile of the number of vehicles passing through all gantry sequences on that day for three consecutive days, and obtain the daily high-frequency gantry pairs of the corresponding day.

5. The traffic congestion perception method based on ETC gantry data as claimed in claim 3, characterized in that: The data of the time window of the activated gantry pair include: the time points of all vehicles passing through the set number of time windows, the passing time T, and the mean value t of the passing time of all vehicles in the set number of time windows; The method for identifying high-frequency gantry pairs with abnormally changing passage time is that the time interval between the earliest and latest timestamps of the passage time set T of the gantry pair exceeds the set time.

6. The traffic congestion perception method based on ETC gantry data according to claim 1, characterized in that: The general congestion fusion perception method based on road section travel time includes the following steps: Traverse each activated high-frequency gantry pair, obtain the travel time of all vehicles in the current time window and the previous set number of time windows, and store them in the current travel time set T′; The travel time of all vehicles in each window of the activated high-frequency gantry pair is perceived using the slope analysis method and the standard deviation analysis method respectively to identify the congested gantry pair sections.

7. The traffic congestion perception method based on ETC gantry data as claimed in claim 6, characterized in that: The slope analysis method is used, with time as the horizontal axis and travel time as the vertical axis, to determine whether a general congestion event occurs based on the slope change, including the following process: Divide the travel time in the current travel time set T′ by the corresponding mean of the travel time to obtain a first multiple set B′; Divide the travel time in the normal travel time set T by the corresponding mean of the travel time to obtain a second multiple set B; Calculate a first slope K1 of the first multiple set B′ with respect to time; Calculate a second slope K2 of the second multiple set B with respect to time; When the difference between the first slope and the second slope exceeds the set value, it is judged as general congestion, and the standard deviation analysis method is performed for the next step of judgment; otherwise, it is judged as no congestion and the process ends; Alternatively, for the gantry pairs that are confirmed to be generally congested by the slope analysis method, the standard deviation analysis method is used to further determine the congestion situation. The specific steps are as follows: Set the standard deviation threshold of the gantry pair's travel time in each time window; Calculate the standard deviation of the travel time of the gantry pair in each time window; If the standard deviation of the travel time is not greater than the standard deviation threshold of the travel time, it can be determined that the road section between the gantry pairs is generally congested; otherwise, the current gantry pair is not congested.

8. The traffic congestion perception method based on ETC gantry data as claimed in claim 1, characterized in that: The zero-pass judgment is used to perceive severe congestion, and the following specific steps are performed for the activated gantry pairs: Step C1: for the current gantry pair, if the number of vehicles passing in the two time windows before the current time window is not less than the set vehicle passing threshold c, then proceed to the next step C2 for judgment; otherwise, the road section between the current gantry pair is not severely congested; Step C2: for the current gantry pair, if the time difference between the last vehicle passing time in the previous time window and the end time of the time window is greater than or equal to the set duration, it is judged to be severely congested; otherwise, the road section between the current gantry pair is not severely congested.

9. Traffic congestion perception system based on ETC gantry data, characterized by: include: Road section division module: It is configured to divide the road sections based on the ETC gantries, match the upstream and downstream gantries in each road section, and determine the bidirectional gantries sequence according to the flow order; Gantry sequence library construction module: It is configured to identify the daily high-frequency gantry pairs according to the vehicle volume, filter the daily high-frequency gantry pairs in all the bidirectional gantry sequences of the road network, build a road network congestion fusion perception gantry sequence library, and remove the invalid gantry pairs in real time for updating; Congestion section identification module: It is configured to activate the daily high-frequency gantry pairs in the road network congestion fusion perception gantry sequence library, and use the slope analysis method and standard deviation analysis method to perform general congestion joint perception based on the road section travel time, and use zero-pass judgment to perceive severe congestion and determine the congested gantry pair sections; Congestion location precise positioning module: It is configured to determine the congested gantry section, perform image recognition through video surveillance data, and determine the specific congested section.

10. An electronic device, characterized in that: The method comprises a memory and a processor and computer instructions stored in the memory and executed on the processor. When the computer instructions are executed by the processor, the steps of the traffic congestion perception method based on ETC gantry data described in any one of claims 1 to 8 are completed.

Images