Tunnel traffic congestion state identification method based on deep learning

Abstract

The invention discloses a tunnel traffic congestion state recognition method based on deep learning, and the method comprises the steps: recognizing a vehicle in a traffic monitoring video, and obtaining the coordinate information of the vehicle in an image; according to the known lane line position and the vehicle coordinate information, determining the lane number of each vehicle; for each lane, a traffic flow statistical area is preset for each lane for traffic flow statistics, and the average traffic flow in unit time is calculated; counting the time for passing vehicles and front vehicles in each lane to arrive at the traffic flow statistical area, and calculating the average time headway in unit time; counting the time required for passing vehicles on each lane to pass through the traffic flow statistical area of each lane, and calculating the average lane passing speed in unit time; weighting coefficients are assigned to the average traffic flow, the average time headway and the average lane passing speed respectively, a tunnel traffic jam index is obtained through weighted calculation, and the tunnel traffic jam state is judged according to the index. Manual intervention is not needed, and automatic detection and recognition are achieved.

Description

Technical Field

[0001] The present invention relates to a method for identifying tunnel traffic congestion status based on deep learning, and belongs to the technical field of tunnel traffic management. Background Art

[0002] Among the existing tunnel traffic congestion status identification methods, it is common to install surveillance cameras in the tunnel and manually monitor to determine the tunnel traffic congestion status, or obtain road congestion status information through map applications. The manual monitoring method requires a lot of manpower, and the standards are not clear, and it relies too much on human attention and judgment. The method of obtaining congestion status through map applications, the information comes from a third party, the reliability is questionable, and the congestion status information obtained is delayed, and the real-time performance is poor.

[0003] The above methods either rely on manual monitoring or are based on big data analysis; these methods do not deeply mine the information of monitoring images, and do not use deep learning and computer vision technology to automatically identify images. At present, technologies such as image target detection and image classification based on deep learning are developing rapidly, but they are basically blank in tunnel traffic monitoring applications. Existing solutions have few functions and are too dependent on manual work. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method for identifying tunnel traffic congestion status based on deep learning, by extracting three traffic congestion factors, weightedly calculating the traffic congestion index, and then judging the congestion status.

[0005] The present invention adopts the following technical solutions to solve the above technical problems:

[0006] A method for identifying tunnel traffic congestion status based on deep learning, comprising the following steps:

[0007] Step 1: Obtain a traffic monitoring video of the tunnel to be identified, identify the vehicle in each frame of the video, and obtain the coordinate information of the vehicle in the image;

[0008] Step 2, determining the lane number of each vehicle based on the known lane line position information and the vehicle coordinate information obtained in step 1;

[0009] Step 3: For each lane, a traffic flow statistics area is pre-set for each lane, traffic flow statistics are performed in the traffic flow statistics area of ​​each lane, and the average traffic flow per unit time is calculated;

[0010] Step 4: For each lane, the time between the vehicles passing through each lane and the preceding vehicle arriving at the traffic flow statistical area is counted, and the headway time of the vehicles passing through each lane is calculated, and then the average headway time per unit time is calculated;

[0011] Step 5: For each lane, count the time required for vehicles passing through each lane to pass through the traffic flow statistical area of ​​each lane, calculate the lane speed, and then calculate the average lane speed per unit time;

[0012] Step 6, assign weight coefficients to the average traffic volume, average headway time and average lane speed per unit time, respectively, obtain the tunnel traffic congestion index by weighted calculation, and judge the tunnel traffic congestion status according to the tunnel traffic congestion index.

[0013] As a preferred solution of the present invention, in step 1, the vehicle in the image is identified by a vehicle target detection algorithm, wherein the vehicle target detection algorithm is a pre-trained YOLOv5 target detection network.

[0014] As a preferred solution of the present invention, the specific process of step 2 is as follows:

[0015] From the position of each vehicle on the image, a ray parallel to the x-axis of the image is emitted to the left, and the number of intersections between the ray and the known lane lines is calculated. The number of intersections is the lane number to which each vehicle belongs.

[0016] As a preferred solution of the present invention, the specific process of step 3 is as follows:

[0017] Preset a traffic flow statistical area for each lane and a time step. For the current time step, determine whether there are vehicles entering the traffic flow statistical area within the current time step according to the vehicles identified in step 1. If yes, determine whether the vehicle entering the traffic flow statistical area is entering the traffic flow statistical area for the first time. If it is the first time, increase the lane passing vehicle count, calculate the increment of the number of passing vehicles within the current time step, and divide the increment by the time step to obtain the average traffic flow per unit time.

[0018] Determine whether the vehicle entering the traffic flow statistical area is entering the traffic flow statistical area for the first time. Specifically, if the vehicle entering the traffic flow statistical area did not enter the traffic flow statistical area in the time step before the current time step, and the vehicle is identified to enter the traffic flow statistical area in the current time step, then it is determined that the vehicle enters the traffic flow statistical area for the first time.

[0019] As a preferred solution of the present invention, in step 4, for each lane, when a vehicle first appears in the traffic flow statistics area, its appearance time is recorded, and for two vehicles passing sequentially, the time difference between the first appearance of the rear vehicle and the front vehicle in the area is calculated, which is the headway of the rear vehicle; the headway of the vehicles in unit time is averaged to obtain the average headway in unit time.

[0020] As a preferred solution of the present invention, in step 6, the weight coefficients of the average traffic flow per unit time, the average headway time and the average lane speed are set according to the traffic statistics data of the actual scene. After the tunnel traffic congestion index is calculated, according to a preset threshold, when the tunnel traffic congestion index is greater than the preset threshold, it is judged that the tunnel is in a congested state.

[0021] A computer device comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the computer program, the steps of the method for identifying tunnel traffic congestion status based on deep learning are implemented.

[0022] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for identifying tunnel traffic congestion status based on deep learning.

[0023] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0024] 1. The present invention uses the image information returned by the monitoring cameras installed in the tunnel to detect the passing vehicles and identify the lanes they are in, and analyzes the vehicle information of all lanes to obtain the traffic flow, headway and speed of each lane, and summarizes the above information to obtain the tunnel traffic congestion status, realizing automatic monitoring and identification without human intervention.

[0025] 2. The present invention extracts three traffic congestion factors accurately and without delay: average vehicle flow, headway, and travel speed, etc., and calculates the traffic congestion index by weighted calculation, and outputs the tunnel traffic congestion status, thereby providing an automated monitoring method for tunnel traffic monitoring, which is more time-saving and labor-saving than manual viewing. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a flow chart of a method for identifying tunnel traffic congestion status based on deep learning of the present invention;

[0027] Figure 2 This is an example of a real-life scenario of a tunnel traffic congestion status recognition method based on deep learning, where (a) is the vehicle detected in the image and (b) is the tunnel congestion status recognized in the image. DETAILED DESCRIPTION

[0028] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0029] like Figure 1 As shown, the present invention proposes a method for identifying tunnel traffic congestion status based on deep learning, and the specific steps are as follows:

[0030] Step S1: Acquire a frame of image in a tunnel traffic monitoring video.

[0031] Step S2: Use a vehicle target detection algorithm based on the YOLOv5 target detection network to identify the vehicle in the image and obtain the coordinate information of the vehicle in the image.

[0032] Step S3: Use the known lane line position information and the image coordinate information obtained by vehicle target detection to identify the lane where each vehicle is located. Specifically:

[0033] A ray parallel to the x-axis of the image is emitted to the left from the position of the vehicle on the image, and the number of intersections between the ray and each lane line is calculated. The number of intersections is the lane number to which the vehicle belongs.

[0034] Step S4: Perform traffic flow statistics in the traffic flow statistics area of ​​each lane. If there is a vehicle that appears for the first time in the area, the statistical value of the number of vehicles passing through the lane is increased, and the average traffic flow per unit time is calculated.

[0035] Specifically: in the traffic statistics area of ​​each lane, search for vehicle target detection results. If a vehicle is detected in the area, read the vehicle status in the area in the previous short period of time. If no vehicle has passed through the area in the previous short period of time, judge that the current vehicle is the first vehicle to appear in this traffic statistics area, and increase the statistical value of the number of vehicles passing through the lane. Then calculate the increment of the number of vehicles passing through per unit time, and output the average traffic flow per unit time.

[0036] Step S5: Count the time between the current vehicle passing through each lane and the preceding vehicle arriving at the traffic flow statistics area, calculate the headway time of each lane, and further calculate the average headway time per unit time.

[0037] Specifically: for each lane, when a vehicle first appears in the traffic flow statistics area, its appearance time is recorded. For two vehicles passing sequentially, the time difference between the first appearance of the rear vehicle and the front vehicle in the area is calculated, which is the headway time of each lane.

[0038] Step S6: Count the time required for the current vehicle to pass through the traffic flow statistical area of ​​each lane, estimate the lane speed, and further calculate the average lane speed per unit time.

[0039] Step S7, comprehensively consider the average vehicle flow rate per unit time, average headway time and average lane speed obtained in steps S4, S5 and S6, weight the above three factors, adjust the weighting coefficient according to the traffic statistics data of the actual scene, calculate the tunnel traffic congestion index, and output the tunnel traffic congestion status. The tunnel traffic congestion status is divided into three states: "smooth", "slow moving" and "congested".

[0040] The method of the present invention was tested with 20,000 road monitoring image data, and the test laboratory conditions were: an Intel i7 9700 CPU, an NVIDIA GTX 2070 GPU with 8G video memory. Figure 2 As shown in (a), the box clearly indicates the vehicle information accurately detected by the vehicle target detection algorithm. Figure 2 As shown in (b), when the tunnel traffic congestion index obtained by weighted calculation is greater than a certain threshold, the threshold is adjusted according to the actual situation. At this time, it is a "congested state". The box in the figure clearly indicates the congested state of the tunnel traffic.

[0041] Based on the same inventive concept, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the aforementioned deep learning-based tunnel traffic congestion status identification method are implemented.

[0042] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the aforementioned deep learning-based tunnel traffic congestion state identification method are implemented.

[0043] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0044] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0045] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0046] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0047] The above embodiments are only for illustrating the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A method for identifying tunnel traffic congestion status based on deep learning, characterized in that: The steps include: Step 1: Obtain a traffic monitoring video of the tunnel to be identified, identify the vehicle in each frame of the video, and obtain the coordinate information of the vehicle in the image; Step 2, determining the lane number of each vehicle based on the known lane line position information and the vehicle coordinate information obtained in step 1; Step 3: For each lane, a traffic flow statistics area is pre-set for each lane, traffic flow statistics are performed in the traffic flow statistics area of ​​each lane, and the average traffic flow per unit time is calculated; Step 4: For each lane, the time between the vehicles passing through each lane and the preceding vehicle arriving at the traffic flow statistical area is counted, and the headway time of the vehicles passing through each lane is calculated, and then the average headway time per unit time is calculated; Step 5: For each lane, count the time required for vehicles passing through each lane to pass through the traffic flow statistical area of ​​each lane, calculate the lane speed, and then calculate the average lane speed per unit time; Step 6, assign weight coefficients to the average traffic volume, average headway time and average lane speed per unit time, respectively, obtain the tunnel traffic congestion index by weighted calculation, and judge the tunnel traffic congestion status according to the tunnel traffic congestion index.

2. The method for identifying tunnel traffic congestion status based on deep learning according to claim 1, characterized in that: In step 1, a vehicle in the image is identified by a vehicle target detection algorithm, wherein the vehicle target detection algorithm is a pre-trained YOLOv5 target detection network.

3. The method for identifying tunnel traffic congestion status based on deep learning according to claim 1, characterized in that: The specific process of step 2 is as follows: From the position of each vehicle on the image, a ray parallel to the x-axis of the image is emitted to the left, and the number of intersections between the ray and the known lane lines is calculated. The number of intersections is the lane number to which each vehicle belongs.

4. The method for identifying tunnel traffic congestion status based on deep learning according to claim 1, characterized in that: The specific process of step 3 is as follows: Preset a traffic flow statistical area for each lane and a time step. For the current time step, determine whether there are vehicles entering the traffic flow statistical area within the current time step according to the vehicles identified in step 1. If yes, determine whether the vehicle entering the traffic flow statistical area is entering the traffic flow statistical area for the first time. If it is the first time, increase the lane passing vehicle count, calculate the increment of the number of passing vehicles within the current time step, and divide the increment by the time step to obtain the average traffic flow per unit time. Determine whether the vehicle entering the traffic flow statistical area is entering the traffic flow statistical area for the first time. Specifically, if the vehicle entering the traffic flow statistical area did not enter the traffic flow statistical area in the time step before the current time step, and the vehicle is identified to enter the traffic flow statistical area in the current time step, then it is determined that the vehicle enters the traffic flow statistical area for the first time.

5. The method for identifying tunnel traffic congestion status based on deep learning according to claim 1, characterized in that: In step 4, for each lane, when a vehicle first appears in the traffic flow statistics area, its appearance time is recorded. For two vehicles passing sequentially, the time difference between the first appearance of the rear vehicle and the front vehicle in the area is calculated, which is the headway time of the rear vehicle. The average headway time of vehicles in unit time is calculated to obtain the average headway time per unit time.

6. The method for identifying tunnel traffic congestion status based on deep learning according to claim 1, characterized in that: In step 6, the weight coefficients of the average traffic volume per unit time, the average headway time and the average lane speed are set according to the traffic statistics data of the actual scene. After the tunnel traffic congestion index is calculated, according to the preset threshold, when the tunnel traffic congestion index is greater than the preset threshold, it is judged that the tunnel is in a congested state.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the computer program, the steps of the method for identifying tunnel traffic congestion status based on deep learning as described in any one of claims 1 to 6 are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for identifying tunnel traffic congestion status based on deep learning as described in any one of claims 1 to 6 are implemented.

Images