An online identification method for traffic section congestion status based on ETC data

Abstract

The present invention discloses an online identification method for the congestion state of a traffic section based on ETC data. By acquiring the ETC data of a gantry, no additional equipment is required. The congestion state of the current traffic section area is determined by the ETC data, and the lane change situation of the vehicle is accurately identified by using the vehicle image data collected by the ETC device in the ETC data. The initial induced value is adjusted by using the lane change situation of each marked vehicle, so that the final congestion induced value of each marked vehicle in the process of passing through the congested area can be accurately understood, thereby judging the congestion-inducing vehicle. The method can accurately identify the slow-speed vehicles that cause congestion by slow-speed driving on the expressway without additionally installing other equipment, and can accurately warn the inducing points of traffic section congestion, provide safe driving support information for drivers, and ensure the driving safety of the majority of drivers.

Description

Technical Field

[0001] The present invention relates to the technical field of big data processing, and in particular to an online identification method for traffic section congestion status based on ETC data. Background Art

[0002] Expressways, or highways for short, are roads designed for high-speed driving of cars. Expressways have different regulations in different countries, regions, eras, and scientific research fields. According to China's Highway Engineering Technical Standards (JTG B01-2014), expressways are multi-lane roads designed for cars to drive in different directions and lanes, with full access control. The average daily design traffic volume of expressways should be more than 15,000 passenger cars, and the design speed can reach 120 kilometers per hour.

[0003] Many people think that speeding can lead to danger when driving on highways, but in fact, driving slowly on highways can also lead to danger. According to statistics, nearly 30% of accidents or congestion on highways are caused by slow driving, and slow driving will inevitably cause road congestion. Relevant personnel and units will monitor the congestion of highway sections, but they often only consider the congestion caused by traffic accidents, locate the location of the traffic accident, and then arrange engineering vehicles for rescue. Even the commonly used map navigation software on the market uses the location of the traffic accident uploaded by the driver to warn of congested sections. Although relevant personnel have proposed to distinguish slow vehicles by monitoring the speed of vehicles, it is impossible to accurately judge that the slow-moving vehicle is the triggering point of road congestion only from the speed of a certain moment or a single vehicle speed, so how to accurately determine the triggering point of road congestion is a technical problem that needs to be solved at present. However, there is no relevant solution strategy for the problem of highway congestion caused by slow driving on highway sections on the market.

[0004] With the development of informatization today, more and more families are buying and using vehicles, and many women and middle-aged and elderly people are also joining in the daily driving. In an era of increasing vehicle sales, vehicle congestion is becoming more and more obvious, and with it comes the problem of safe vehicle driving. How to analyze the road congestion caused by slow driving on the expressway, accurately determine the slow vehicles that cause road congestion and conduct early warning navigation, is a difficult problem for safe driving on the expressway. In the prior art, a gantry is set up in each section of the expressway, and an ETC device for photo-taking and charging is installed on the gantry. Every time a vehicle passes through a gantry, the ETC device above will take a photo and charge according to the vehicle model. At the same time, the ETC device will exchange data with the on-board OBU on the vehicle. However, with the increase in costs, how to use the existing ETC equipment to solve the above-mentioned safe driving problems without installing other equipment has also become the focus of engineers.

[0005] Therefore, there is an urgent need on the market for an online identification strategy for traffic section congestion status based on ETC data to solve the technical problem that the existing technology is unable to identify and warn slow vehicles that cause congestion on highways due to slow speed. Summary of the invention

[0006] The present invention provides an online identification method for the congestion status of traffic section intervals based on ETC data. Without installing other additional equipment, the existing ETC data is used to accurately identify slow vehicles that cause congestion due to slow driving on expressway sections. Accurate early warning can be given for the inducing points of congestion in traffic section intervals, and safe driving support information can be provided to drivers, thereby ensuring the driving safety of the majority of drivers.

[0007] In order to solve the above technical problems, an embodiment of the present invention provides an online identification method for traffic section congestion status based on ETC data, comprising:

[0008] Acquire ETC data passing through the gantry in real time; wherein the ETC data includes vehicle image data and ETC interaction data;

[0009] According to the ETC interaction data, real-time statistics are collected on the vehicle flow through the gantry within a preset unit time, and the average speed of the vehicles passing through the preset unit time is determined according to the vehicle image data, and the congestion value of the current traffic section area is determined according to the average speed and the vehicle flow;

[0010] When the congestion value reaches a preset threshold, the current traffic section area is determined to be in a congested state, and all vehicles passing through the gantry within the preset unit time are defined as suspected slow vehicles according to the vehicle image data;

[0011] Extracting the vehicle image data formed within the preset unit time length to generate vehicle driving dynamic data, inputting the vehicle driving dynamic data into a preset congestion area recognition model, identifying vehicles forming a congestion section, and determining and outputting the marked vehicles forming the congestion section from the suspected slow vehicles; at the same time, assigning a same initial induced value to all the marked vehicles;

[0012] Extracting vehicle driving images containing marked vehicles from the vehicle image data, determining driving lanes of all marked vehicles, and determining lane change conditions of each marked vehicle in the vehicle driving image based on the driving lanes of the marked vehicles;

[0013] Adjusting the initial induced value of each marked vehicle according to the lane change situation of the marked vehicle to obtain a congestion induced value;

[0014] The first N marked vehicles with the largest congestion inducing values ​​are determined as congestion inducing vehicles, and the current lane and vehicle information of the congestion inducing vehicles are extracted as the inducing points of congestion in the current traffic section, and sent to the traffic command center for early warning.

[0015] As a preferred solution, after the vehicle flow passing through the gantry within a preset unit time length is counted in real time according to the ETC interaction data, the method further includes:

[0016] According to the ETC interaction data, filtering the vehicles for which ETC interaction statistics have been performed in the vehicle image data, and obtaining the vehicles without the vehicle-mounted OBU installed in the vehicle image data;

[0017] The vehicle flow is updated according to the number of vehicles without on-board OBU installed that pass through the gantry within a preset unit time.

[0018] As a preferred solution, the step of determining the average speed of vehicles passing through the preset unit time period according to the vehicle image data specifically includes:

[0019] Determine multiple monitoring time points within the preset unit time length, extract vehicle images corresponding to the monitoring time points from the vehicle image data, and arrange the vehicle images in sequence according to the sequence of the monitoring time points; wherein the monitoring time points include the initial time point and the final time point of the preset unit time length;

[0020] Identify vehicle features in the first frame of the vehicle image, divide the first frame of the vehicle image into a plurality of driving areas and determine the reference point position in each driving area respectively; wherein each driving area contains at least one vehicle;

[0021] Identify the vehicle features in the next adjacent vehicle image frame, determine the reference point position of each driving area in the next vehicle image frame, and calculate the average speed of each reference point in the two adjacent vehicle image frames according to the distance formed by the same reference point in the two adjacent vehicle image frames and the time difference between the two vehicle image frames;

[0022] The average speed of each reference point in two adjacent frames of vehicle images is counted, and the average speed between two adjacent frames of vehicle images is calculated. By calculating the average of the average speeds between all two adjacent frames of vehicle images, the average speed of the vehicles passing within a preset unit time length is obtained.

[0023] As a preferred solution, the calculation formula of the congestion value is:

[0024]

[0025] Where: W is the congestion value; I is the vehicle flow, I0 is the preset flow value; V is the average speed of vehicles passing through the preset unit time, V0 is the preset speed value; k1 and k2 are both constants;

[0026] in,

[0027] Among them, V j is the average speed between two adjacent frames of vehicle images, j refers to two adjacent frames of vehicle images, m is the total number of vehicle images - 1; h ij is the distance between the i-th reference point in two adjacent vehicle images, i refers to the reference point, n is the total number of reference points, and t j is the time difference between two adjacent frames of vehicle images.

[0028] As a preferred solution, the step of extracting the vehicle image data formed within the preset unit time to generate vehicle driving dynamic data specifically includes:

[0029] Extracting the vehicle images collected within the preset unit time length from the vehicle image data as input images;

[0030] Identify vehicle features in each frame of input image, mark each vehicle, establish the same spatial coordinate system in each frame of input image, and determine the position of each marked vehicle in the spatial coordinate system;

[0031] According to the position of each marked vehicle in the spatial coordinate system, calculating the moving position of each marked vehicle in two adjacent frames of input images and determining its acceleration value;

[0032] Mark each vehicle in each frame of input image with the acceleration value between the current input image and the previous input image to obtain an acceleration image set;

[0033] All image frames in the acceleration image set are sequentially sorted and combined in chronological order to obtain vehicle driving dynamic data.

[0034] As a preferred solution, the process of constructing the preset congestion area identification model specifically includes:

[0035] Acquire a historical vehicle image; wherein the traffic section area displayed in the historical vehicle image is in a congested state;

[0036] Obtaining the acceleration value of each vehicle in the historical vehicle image within a preset unit time length, and marking the acceleration value at the corresponding vehicle position in the historical vehicle image;

[0037] Determining a plurality of inducing points that cause congestion in the traffic section area in the historical vehicle image, and associating the plurality of inducing points with the marked vehicle positions respectively to form a congestion inducing network image;

[0038] An initial recognition model is established through a neural network algorithm, and the congestion-inducing network image is input into the initial recognition model for training. When the number of training times and the training accuracy reach preset values, the training is completed to obtain a congested area recognition model.

[0039] As a preferred solution, the step of extracting the vehicle driving image containing the marked vehicle from the vehicle image data and determining the driving lanes of all the marked vehicles specifically includes:

[0040] The vehicle driving images containing the marked vehicles are sequentially sorted in chronological order, and the lane lines in the vehicle driving images are identified to determine the positions of the lane lines;

[0041] Perform feature recognition on the marked vehicle, extract the two highlight points at the leftmost and rightmost ends of the marked vehicle respectively, calculate the distance values ​​between the two highlight points and the positions of each lane line in the horizontal direction respectively, and determine the first lane line position corresponding to the minimum distance value from the highlight point at the leftmost end, and the second lane line position corresponding to the minimum distance value from the highlight point at the rightmost end;

[0042] The lane between the first lane line position and the second lane line position is determined as a driving lane for the marked vehicle.

[0043] As a preferred solution, the step of determining the lane change situation of each marked vehicle in the vehicle driving image according to the driving lane of the marked vehicle specifically includes:

[0044] Determine the driving lane of the marked vehicle in each frame of the vehicle driving image, and when it is determined that the driving lane of the marked vehicle in the next frame of the vehicle driving image is different from that in the previous frame, determine that the marked vehicle has changed lanes; at the same time, determine the target preceding vehicle of the marked vehicle in the previous frame of the vehicle driving image;

[0045] After determining that the marked vehicle has changed lanes, whether it has overtaken the target vehicle in front within a preset first lane change time period, if so, determining that the lane change of the marked vehicle is a valid lane change; otherwise, determining whether the marked vehicle sends a lane change again within a preset second lane change time period;

[0046] When it is determined that the marked vehicle sends a lane change again within the preset second lane change time period, determining whether the marked vehicle changes lanes to the original driving lane, if yes, determining that the two consecutive lane changes of the marked vehicle are invalid lane changes; otherwise, determining that the two consecutive lane changes of the marked vehicle are pending lane changes;

[0047] When it is determined that the marked vehicle does not send a lane change again within the preset second lane change time period, the current lane change of the marked vehicle is determined to be a pending lane change.

[0048] As a preferred solution, the step of adjusting the initial induced value of each marked vehicle according to the lane change situation of each marked vehicle to obtain the congestion induced value specifically includes:

[0049] When it is determined that the lane change of the marked vehicle is a valid lane change, the initial induced value of the marked vehicle is subtracted by 1, and the initial induced value of the target front vehicle is increased by 1;

[0050] When it is determined that the lane change of the marked vehicle is an invalid lane change, the initial induced value of the marked vehicle is not processed, and the initial induced value of the target front vehicle is reduced by 1;

[0051] When it is determined that the lane change situation of the marked vehicle is a pending lane change, the initial induced value of the marked vehicle is increased by 1, and the initial induced value of the target front vehicle is not processed;

[0052] After the lane changing conditions of each marked vehicle in all vehicle driving images are traversed and adjusted, the congestion inducing value corresponding to each marked vehicle is obtained.

[0053] An embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program; wherein, when the computer program is running, it controls the device where the computer-readable storage medium is located to execute the online identification method of the congestion status of a traffic section based on ETC data as described in any one of the above.

[0054] An embodiment of the present invention also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the online identification method for the congestion status of a traffic section based on ETC data as described in any one of the above when executing the computer program.

[0055] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0056] The technical solution of the present invention obtains the ETC data of the gantry without the need to install other additional equipment. The congestion status of the current traffic section area is determined by the ETC data, and the vehicle image data collected by the ETC device in the ETC data is used to accurately identify the lane change of the vehicle. The initial induced value is adjusted according to the lane change of each marked vehicle, so that the final congestion induced value of each marked vehicle in the process of passing through the congested area can be accurately understood, thereby judging the congestion-inducing vehicle; the technical problem that the prior art cannot identify and warn slow-speed vehicles that cause congestion due to slow driving on expressway sections, and the existing ETC data can be used to accurately identify slow-speed vehicles that cause congestion due to slow driving on expressway sections without installing other additional equipment, and accurate warnings can be given to the inducing points of congestion in the traffic section, providing safe driving support information for drivers, and ensuring the driving safety of the majority of drivers. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 : A flowchart of a method for online identification of traffic section congestion status based on ETC data provided by an embodiment of the present invention;

[0058] Figure 2 : A structural diagram of an embodiment of a terminal device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] Embodiment 1

[0061] Please refer to Figure 1 , is a flowchart of a method for online identification of traffic section congestion status based on ETC data provided by an embodiment of the present invention, including steps 101 to 107, each of which is specifically as follows:

[0062] Step 101, acquiring ETC data passing through the gantry in real time; wherein the ETC data includes vehicle image data and ETC interaction data.

[0063] Specifically, in order to achieve the goal of not installing additional equipment, we collect vehicle data on the highway section through the ETC equipment normally set on the gantry. It can be understood that the ETC data in this step includes the vehicle image data taken by the ETC equipment of the vehicle traveling on the road, and the ETC equipment and the vehicle-mounted OBU installed on the traveling vehicle to exchange data to obtain the ETC interaction data of vehicle information.

[0064] Step 102, based on the ETC interaction data, real-time statistics are collected on the vehicle flow passing through the gantry within a preset unit time, and the average speed of the vehicles passing through the preset unit time is determined based on the vehicle image data, and the congestion value of the current traffic section area is determined based on the average speed and the vehicle flow.

[0065] Specifically, this step first uses ETC interaction data to determine the traffic volume. We can use the interaction between the ETC device and the vehicle-mounted OBU to determine how many vehicles pass through the current gantry within a preset unit time, thereby determining the traffic volume. In addition, we also need to consider two aspects. On the one hand, how to use vehicle image data to determine the average speed of the vehicle; on the other hand, how to use the average speed and traffic volume to calculate the congestion value to determine the congestion status of the current traffic section area. The following will explain in detail.

[0066] In the first aspect of the present embodiment, the step of determining the average speed of the vehicle passing through the preset unit time length according to the vehicle image data specifically includes: step 1021, determining multiple monitoring time points within the preset unit time length, extracting the vehicle images corresponding to the monitoring time points from the vehicle image data, and arranging the vehicle images in sequence according to the sequence of the monitoring time points; wherein the monitoring time points include the initial time point and the final time point of the preset unit time length. Step 1022, identifying the vehicle features in the first frame of the vehicle image, dividing the first frame of the vehicle image into multiple driving areas and determining the reference point position in each driving area respectively; wherein each driving area contains at least one vehicle. Step 1023, identifying the vehicle features in the next adjacent frame of the vehicle image, determining the reference point position of each driving area in the next frame of the vehicle image, and calculating the average speed formed by each reference point in the two adjacent frames of the vehicle image according to the distance formed by the same reference point in the two adjacent frames of the vehicle image and the time difference between the two frames of the vehicle image. Step 1024, counting the average speed of each reference point in two adjacent frames of vehicle images, calculating the average speed between two adjacent frames of vehicle images, and calculating the average of the average speeds between all two adjacent frames of vehicle images to obtain the average speed of vehicles passing within a preset unit time length.

[0067] Specifically, without adding other equipment, how to use the only ETC equipment to determine the average speed of the vehicle is actually a difficult problem. Because the equipment used to determine the speed of the vehicle in the prior art is a monitoring device, the ETC device itself does not have a speed measurement function. The monitoring equipment on the expressway is not installed on every gantry, so not all expressway areas can measure the speed of the vehicle. In addition, the speed measurement section intervals on the expressway are unevenly set, so when using the ETC equipment for data collection, it is impossible to use the existing monitoring equipment to measure the speed of the vehicle simultaneously. Therefore, we need to develop a breakthrough strategy to complete the calculation of the average speed of the vehicle using only ETC data. In this step, we first set the monitoring time point, and use the vehicle image data collected in the previous step to extract the vehicle images at these monitoring time points. Note that at this time, we need to extract the vehicle images at the initial time point and the final time point in the preset unit time length to make the subsequent calculation of the average speed of the vehicle more accurate. The collected vehicle image starts with the first frame image and begins to identify the vehicle in the image. Since some vehicles may travel in the same area during driving, we need to divide the vehicle image into multiple driving areas, which are formed by multiple vehicles driving. Therefore, these driving areas contain at least one vehicle, and we use a vehicle in these driving areas as the reference point position, so we also need to determine the reference point position in the step. It can be understood that there are multiple variations in the determination of the reference point position. We only need to determine the reference point of the driving area, and then use the position movement of these reference points to determine the position movement of the corresponding driving area. Therefore, the choice of the reference point position is not limited here. Next, we only need to calculate the time point corresponding to the position movement, and the average speed can be obtained by calculating the ratio.

[0068] Specifically, after we calculate the average speed and traffic volume, we can use the calculation formula of the congestion value to calculate the congestion value. Studies have shown that the relationship between traffic volume and average speed can accurately reflect the congestion of the road. Therefore, in the second aspect of this embodiment, the calculation formula of the congestion value is:

[0069]

[0070] Where: W is the congestion value; I is the vehicle flow, I0 is the preset flow value; V is the average speed of vehicles passing through the preset unit time, V0 is the preset speed value; k1 and k2 are both constants;

[0071] in,

[0072] Among them, V jis the average speed between two adjacent frames of vehicle images, j refers to two adjacent frames of vehicle images, m is the total number of vehicle images - 1; h ij is the distance between the i-th reference point in two adjacent vehicle images, i refers to the reference point, n is the total number of reference points, and t j is the time difference between two adjacent frames of vehicle images.

[0073] Furthermore, in another embodiment, after the vehicle flow passing through the gantry within a preset unit time is counted in real time according to the ETC interaction data, it also includes: filtering the vehicles for which ETC interaction statistics have been performed in the vehicle image data according to the ETC interaction data, and obtaining vehicles without on-board OBU installed in the vehicle image data; and updating the vehicle flow according to the number of vehicles without on-board OBU installed that pass through the gantry within the preset unit time.

[0074] Specifically, in order to judge the congestion on the highway more accurately. Considering that a very small number of car owners still do not install the on-board OBU, they cannot interact with the ETC equipment on the gantry. Therefore, in order to solve the possible data missing problem in the ETC interaction data, we use vehicle image data to identify the vehicle. First, we filter the ETC interaction data and the vehicles in the statistics, which can save a lot of image calculations to improve processing efficiency. Then we filter out the vehicles that have not interacted, and use the image recognition function to identify and extract the vehicle information of these vehicles (such as vehicle model, license plate number, etc.) to supplement the traffic data.

[0075] Step 103, when the congestion value reaches a preset threshold, the current traffic section area is determined to be in a congested state, and based on the vehicle image data, all vehicles passing through the gantry within the preset unit time are defined as suspected slow vehicles.

[0076] Specifically, after determining the congestion, we can accurately determine the vehicles that passed through the gantry during these time periods based on the time points corresponding to the congestion. We first define these vehicles as suspected slow vehicles and then accurately determine them later.

[0077] Step 104, extracting the vehicle image data formed within the preset unit time length, generating vehicle driving dynamics data, inputting the vehicle driving dynamics data into a preset congestion area recognition model, identifying the vehicles forming the congestion section, determining and outputting the marked vehicles forming the congestion section among the suspected slow vehicles; at the same time, assigning a same initial induced value to all the marked vehicles.

[0078] Specifically, we extract vehicle images corresponding to the duration of congestion, and use these image data to generate input data for the congestion area recognition model, namely, vehicle driving dynamic data. The congestion area recognition model here is used to recognize the input image, and then mark the vehicles that form the congestion interval in the input image. This will be explained in detail below.

[0079] In the first aspect of the present embodiment, the step of extracting the vehicle image data formed within the preset unit time length to generate the vehicle driving dynamic data specifically includes: step 10411, extracting the vehicle image collected within the preset unit time length from the vehicle image data as the input image. Step 10412, identifying the vehicle features in each frame of the input image, marking each vehicle, and establishing the same spatial coordinate system in each frame of the input image to determine the position of each marked vehicle in the spatial coordinate system. Step 10413, according to the position of each marked vehicle in the spatial coordinate system, calculating the moving position of each marked vehicle in two adjacent frames of input images and determining its acceleration value. Step 10414, marking the acceleration value between the current input image and the previous input image for each vehicle in each frame of the input image, and obtaining an acceleration image set. Step 10415, sequentially sorting and combining all the image frames in the acceleration image set in chronological order to obtain the vehicle driving dynamic data.

[0080] Specifically, before we input the image into the congested area recognition model, we need to process the image for subsequent model recognition. In the process of model recognition, the principle is that the recognition model determines the inducing point in the input image by the relationship between the vehicle acceleration in the input image and the true inducing point. Therefore, we need to mark the acceleration value of the input image. We establish a spatial coordinate system, mark the moving position of the vehicle between each frame of the image, and combine the time difference between the two frames of the image to calculate the acceleration value. Mark the obtained acceleration value in the image, sort and combine the marked images in sequence to form a data stream.

[0081] In the second aspect of the present embodiment, the construction process of the preset congestion area recognition model specifically includes: Step 10421, obtaining a historical vehicle image; wherein the traffic section area displayed in the historical vehicle image is in a congested state. Step 10422, obtaining the acceleration value of each vehicle in the historical vehicle image within a preset unit time length, and marking the acceleration value on the corresponding vehicle position in the historical vehicle image. Step 10423, determining multiple inducing points that cause congestion in the traffic section area in the historical vehicle image, and associating the multiple inducing points with the marked vehicle positions respectively to form a congestion-inducing network image. Step 10424, establishing an initial recognition model through a neural network algorithm, and inputting the congestion-inducing network image into the initial recognition model for training, until the number of training times and the training accuracy reach the preset values, the training is completed, and the congestion area recognition model is obtained.

[0082] Specifically, the above content is the construction process of a specific congestion area recognition model. Through the above construction strategy, we can obtain an accurate recognition model for accurate recognition between vehicle acceleration and inducing points in the input image, thereby determining the inducing points in the input image to help filter a large number of suspected slow vehicles, and obtain the marked vehicles that form the congestion interval among the suspected slow vehicles and output them. At the same time, we assign the same initial inducing value to all marked vehicles to facilitate the subsequent determination of the real inducing vehicles.

[0083] Step 105, extracting the vehicle driving images containing the marked vehicles from the vehicle image data, determining the driving lanes of all the marked vehicles, and determining the lane change situation of each marked vehicle in the vehicle driving image based on the driving lanes of the marked vehicles.

[0084] Specifically, this step involves two aspects: one aspect is how to determine the vehicle's driving lane; the other aspect is how to use the vehicle's driving lane to determine the vehicle's lane change situation, so as to facilitate the determination of the inducing vehicle in the subsequent process.

[0085] In the first aspect of this embodiment, the step of extracting the vehicle driving images containing the marked vehicles from the vehicle image data and determining the driving lanes of all the marked vehicles specifically includes: step 10511, sorting the vehicle driving images containing the marked vehicles in chronological order, identifying the lane lines in the vehicle driving images, and determining the lane line positions. Step 10512, performing feature recognition on the marked vehicles, extracting the two leftmost and rightmost highlight points in the marked vehicles respectively, calculating the horizontal distance values ​​between the two highlight points and each lane line position respectively, and determining the first lane line position corresponding to the minimum distance value from the leftmost highlight point, and the second lane line position corresponding to the minimum distance value from the rightmost highlight point. Step 10513, determining the lane between the first lane line position and the second lane line position as the driving lane of the marked vehicle.

[0086] Specifically, we first identify the lane lines. Using the identification marks of the vehicle in the vehicle driving image, we need to determine which lane the marked vehicle is closest to. Because sometimes the vehicle does not travel between the lane lines, it may cross the lane or press the line. In this case, we need to process the image by marking the two highlights generated by the vehicle in the image. The two highlights on the left and right are collected here. It should be understood that the leftmost and rightmost highlights here can be the positions of the left and right headlights of the vehicle, which facilitates image recognition and convenient calculation of the distance between the highlight and the lane line. Then we determine the lane line with the smallest distance value. Determine the driving lane of the marked vehicle based on the lane between the two lane lines.

[0087] In the second aspect of the present embodiment, the step of determining the lane change of each marked vehicle in the vehicle driving image according to the driving lane of the marked vehicle specifically includes: step 10521, determining the driving lane of the marked vehicle in each frame of the vehicle driving image, and when it is determined that the driving lane of the marked vehicle in the next frame of the vehicle driving image is different from that in the previous frame, determining that the marked vehicle has changed lanes; at the same time, determining the target front vehicle of the marked vehicle in the previous frame of the vehicle driving image. Step 10522, determining whether the marked vehicle has surpassed the target front vehicle within a preset first lane change time period after determining that the lane change has occurred, if it has surpassed, determining that the lane change of the marked vehicle is a valid lane change; otherwise, determining whether the marked vehicle sends a lane change again within a preset second lane change time period. Step 10523, when it is determined that the marked vehicle sends a lane change again within the preset second lane change time period, determining whether the marked vehicle has changed lanes to the original driving lane, if so, determining that the two consecutive lane changes of the marked vehicle are invalid lane changes; otherwise, determining that the two consecutive lane changes of the marked vehicle are pending lane changes. Step 10524: When it is determined that the marked vehicle does not send a lane change request again within the preset second lane change time period, the lane change of the marked vehicle is determined to be a pending lane change.

[0088] Specifically, we divide the lane changing situations here into three cases: the first is that the marked vehicle changes lanes, and after changing lanes within a certain period of time, it overtakes the vehicle in front of it in the original lane. This is an effective lane change. The second is that the marked vehicle does not overtake the vehicle in front after changing lanes, but changes lanes again within a certain period of time and changes back to the original lane. In this case, it is considered that the road conditions of the marked vehicle are worse after the lane change, while the vehicle in front of it in the original lane actually has no problem, reducing the possibility of the vehicle in front being the triggering point. The third is that although the marked vehicle changes lanes twice, it does not change back to the original lane, but changes to the third lane; or, after changing lanes, the vehicle neither overtakes the vehicle in front nor changes lanes again. In this case, we believe that the marked vehicle may be an inducing point for slow driving.

[0089] Step 106 , adjusting the initial induced value of each marked vehicle according to the lane change situation of the marked vehicle to obtain a congestion induced value.

[0090] In this embodiment, step 106 specifically includes: step 1061, when it is determined that the lane change situation of the marked vehicle is a valid lane change, the initial induced value of the marked vehicle is subtracted by 1, and the initial induced value of the target preceding vehicle is added by 1. Step 1062, when it is determined that the lane change situation of the marked vehicle is an invalid lane change, the initial induced value of the marked vehicle is not processed, and the initial induced value of the target preceding vehicle is subtracted by 1. Step 1063, when it is determined that the lane change situation of the marked vehicle is a pending lane change, the initial induced value of the marked vehicle is added by 1, and the initial induced value of the target preceding vehicle is not processed. Step 1064, after the lane change situation of each marked vehicle in all vehicle driving images is traversed and adjusted, the congestion induced value corresponding to each marked vehicle is obtained.

[0091] Specifically, research shows that through the above adjustment strategy for the initial triggering value, the contribution of each marked vehicle to the congestion state during driving can be accurately determined, thereby ultimately determining the congestion triggering value of each marked vehicle. It can be understood that the higher the congestion triggering value, the higher the possibility that the marked vehicle is an triggering point.

[0092] Step 107, determine the first N marked vehicles with the largest congestion inducing values ​​as congestion inducing vehicles, extract the current lane and vehicle information of the congestion inducing vehicles as the congestion inducing points of the current traffic section, and send them to the traffic command center for early warning.

[0093] Specifically, we can select the first N marked vehicles with the largest congestion inducing values ​​as the vehicles causing congestion according to actual needs, and complete the early warning. It can be understood that the vehicle information here can include information such as vehicle model and license plate number, which can be obtained through ETC interactive data or image recognition.

[0094] The technical solution of the present invention obtains the ETC data of the gantry without the need to install other additional equipment. The congestion status of the current traffic section area is determined by the ETC data, and the vehicle image data collected by the ETC device in the ETC data is used to accurately identify the lane change of the vehicle. The initial induced value is adjusted according to the lane change of each marked vehicle, so that the final congestion induced value of each marked vehicle in the process of passing through the congested area can be accurately understood, thereby judging the congestion-inducing vehicle; the technical problem that the prior art cannot identify and warn slow-speed vehicles that cause congestion due to slow driving on expressway sections, and the existing ETC data can be used to accurately identify slow-speed vehicles that cause congestion due to slow driving on expressway sections without installing other additional equipment, and accurate warnings can be given to the inducing points of congestion in the traffic section, providing safe driving support information for drivers, and ensuring the driving safety of the majority of drivers.

[0095] Embodiment 2

[0096] An embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program; wherein, when the computer program is running, it controls the device where the computer-readable storage medium is located to execute the online identification method of traffic section congestion status based on ETC data as described in any of the above embodiments.

[0097] Embodiment 3

[0098] Please refer to Figure 2 , is a structural diagram of an embodiment of a terminal device provided in an embodiment of the present invention, wherein the terminal device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the online identification method for the congestion status of a traffic section based on ETC data as described in any of the above embodiments when executing the computer program.

[0099] Preferably, the computer program can be divided into one or more modules / units (such as computer program, computer program), and the one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments that can complete specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device.

[0100] The processor may be a central processing unit (CPU), or other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may be any conventional processor. The processor is the control center of the terminal device, and various parts of the terminal device are connected by various interfaces and lines.

[0101] The memory mainly includes a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function, etc., and the data storage area can store related data, etc. In addition, the memory can be a high-speed random access memory, or a non-volatile memory, such as a plug-in hard disk, a smart memory card (SmartMed ia Card, SMC), a secure digital (Secure Digital, SD) card, and a flash card (Flash Card), etc., or the memory can also be other volatile solid-state storage devices.

[0102] It should be noted that the above-mentioned terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that the above-mentioned terminal device is merely an example and does not constitute a limitation on the terminal device. It may include more or fewer components, or a combination of certain components, or different components.

[0103] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. An online identification method for traffic section congestion status based on ETC data, characterized in that: include: Acquire ETC data passing through the gantry in real time; wherein the ETC data includes vehicle image data and ETC interaction data; According to the ETC interaction data, real-time statistics are collected on the vehicle flow through the gantry within a preset unit time, and the average speed of the vehicles passing through the preset unit time is determined according to the vehicle image data, and the congestion value of the current traffic section area is determined according to the average speed and the vehicle flow; When the congestion value reaches a preset threshold, the current traffic section area is determined to be in a congested state, and all vehicles passing through the gantry within the preset unit time are defined as suspected slow vehicles according to the vehicle image data; Extracting the vehicle image data formed within the preset unit time length to generate vehicle driving dynamic data, inputting the vehicle driving dynamic data into a preset congestion area recognition model, identifying vehicles forming a congestion section, and determining and outputting the marked vehicles forming the congestion section from the suspected slow vehicles; at the same time, assigning a same initial induced value to all the marked vehicles; Extracting vehicle driving images containing marked vehicles from the vehicle image data, determining driving lanes of all marked vehicles, and determining lane change conditions of each marked vehicle in the vehicle driving image based on the driving lanes of the marked vehicles; Adjusting the initial induced value of each marked vehicle according to the lane change situation of the marked vehicle to obtain a congestion induced value; Determine the first N marked vehicles with the largest congestion inducing values ​​as congestion inducing vehicles, extract the current lane and vehicle information of the congestion inducing vehicles as the congestion inducing points of the current traffic section, and send them to the traffic command center for early warning; The step of determining the lane change situation of each marked vehicle in the vehicle driving image according to the driving lane of the marked vehicle specifically includes: Determine the driving lane of the marked vehicle in each frame of the vehicle driving image, and when it is determined that the driving lane of the marked vehicle in the next frame of the vehicle driving image is different from that in the previous frame, determine that the marked vehicle has changed lanes; at the same time, determine the target preceding vehicle of the marked vehicle in the previous frame of the vehicle driving image; After determining that the marked vehicle has changed lanes, whether it has overtaken the target vehicle in front within a preset first lane change time period, if so, determining that the lane change of the marked vehicle is a valid lane change; otherwise, determining whether the marked vehicle sends a lane change again within a preset second lane change time period; When it is determined that the marked vehicle sends a lane change again within the preset second lane change time period, determining whether the marked vehicle changes lanes to the original driving lane, if yes, determining that the two consecutive lane changes of the marked vehicle are invalid lane changes; otherwise, determining that the two consecutive lane changes of the marked vehicle are pending lane changes; When it is determined that the marked vehicle does not send a lane change again within the preset second lane change time period, determining the lane change of the marked vehicle as a pending lane change; The step of adjusting the initial induced value of each marked vehicle according to the lane change situation of each marked vehicle to obtain the congestion induced value specifically includes: When it is determined that the lane change of the marked vehicle is a valid lane change, the initial induced value of the marked vehicle is subtracted by 1, and the initial induced value of the target front vehicle is increased by 1; When it is determined that the lane change of the marked vehicle is an invalid lane change, the initial induced value of the marked vehicle is not processed, and the initial induced value of the target front vehicle is reduced by 1; When it is determined that the lane change situation of the marked vehicle is a pending lane change, the initial induced value of the marked vehicle is increased by 1, and the initial induced value of the target front vehicle is not processed; After the lane changing conditions of each marked vehicle in all vehicle driving images are traversed and adjusted, the congestion inducing value corresponding to each marked vehicle is obtained.

2. The method for online identification of traffic section congestion status based on ETC data according to claim 1, characterized in that: After the vehicle flow passing through the gantry within a preset unit time length is counted in real time according to the ETC interaction data, the method further includes: According to the ETC interaction data, filtering the vehicles for which ETC interaction statistics have been performed in the vehicle image data, and obtaining the vehicles without the vehicle-mounted OBU installed in the vehicle image data; The vehicle flow is updated according to the number of vehicles without on-board OBU installed that pass through the gantry within a preset unit time.

3. The method for online identification of traffic section congestion status based on ETC data as claimed in claim 1, characterized in that: The step of determining the average speed of vehicles passing through the preset unit time period according to the vehicle image data specifically includes: Determine multiple monitoring time points within the preset unit time length, extract vehicle images corresponding to the monitoring time points from the vehicle image data, and arrange the vehicle images in sequence according to the sequence of the monitoring time points; wherein the monitoring time points include the initial time point and the final time point of the preset unit time length; Identify vehicle features in the first frame of the vehicle image, divide the first frame of the vehicle image into a plurality of driving areas and determine the reference point position in each driving area respectively; wherein each driving area contains at least one vehicle; Identify the vehicle features in the next adjacent vehicle image frame, determine the reference point position of each driving area in the next vehicle image frame, and calculate the average speed of each reference point in the two adjacent vehicle image frames according to the distance formed by the same reference point in the two adjacent vehicle image frames and the time difference between the two vehicle image frames; The average speed of each reference point in two adjacent frames of vehicle images is counted, and the average speed between two adjacent frames of vehicle images is calculated. By calculating the average of the average speeds between all two adjacent frames of vehicle images, the average speed of the vehicles passing within a preset unit time length is obtained.

4. The method for online identification of traffic section congestion status based on ETC data as claimed in claim 1, characterized in that: The step of extracting the vehicle image data formed within the preset unit time to generate vehicle driving dynamic data specifically includes: Extracting the vehicle images collected within the preset unit time length from the vehicle image data as input images; Identify vehicle features in each frame of input image, mark each vehicle, establish the same spatial coordinate system in each frame of input image, and determine the position of each marked vehicle in the spatial coordinate system; According to the position of each marked vehicle in the spatial coordinate system, calculating the moving position of each marked vehicle in two adjacent frames of input images and determining its acceleration value; Mark each vehicle in each frame of input image with the acceleration value between the current input image and the previous input image to obtain an acceleration image set; All image frames in the acceleration image set are sequentially sorted and combined in chronological order to obtain vehicle driving dynamic data.

5. The method for online identification of traffic section congestion status based on ETC data as claimed in claim 4, characterized in that: The process of constructing the preset congestion area identification model specifically includes: Acquire a historical vehicle image; wherein the traffic section area displayed in the historical vehicle image is in a congested state; Obtaining the acceleration value of each vehicle in the historical vehicle image within a preset unit time length, and marking the acceleration value at the corresponding vehicle position in the historical vehicle image; Determining a plurality of inducing points that cause congestion in the traffic section area in the historical vehicle image, and associating the plurality of inducing points with the marked vehicle positions respectively to form a congestion inducing network image; An initial recognition model is established through a neural network algorithm, and the congestion-inducing network image is input into the initial recognition model for training. When the number of training times and the training accuracy reach preset values, the training is completed to obtain a congested area recognition model.

6. The method for online identification of traffic section congestion status based on ETC data according to claim 1, characterized in that: The step of extracting the vehicle driving image containing the marked vehicle from the vehicle image data and determining the driving lanes of all the marked vehicles specifically includes: The vehicle driving images containing the marked vehicles are sequentially sorted in chronological order, and the lane lines in the vehicle driving images are identified to determine the positions of the lane lines; Perform feature recognition on the marked vehicle, extract the two highlight points at the leftmost and rightmost ends of the marked vehicle respectively, calculate the distance values ​​between the two highlight points and the positions of each lane line in the horizontal direction respectively, and determine the first lane line position corresponding to the minimum distance value from the highlight point at the leftmost end, and the second lane line position corresponding to the minimum distance value from the highlight point at the rightmost end; The lane between the first lane line position and the second lane line position is determined as a driving lane for the marked vehicle.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program; wherein, when the computer program is running, it controls the device where the computer-readable storage medium is located to execute the online identification method of the congestion state of a traffic section based on ETC data as described in any one of claims 1-6.

8. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When executing the computer program, the processor implements the online identification method for the congestion state of a traffic section based on ETC data as described in any one of claims 1 to 6.

Images