Entrance and exit anti-blocking vehicle management system based on camera and laser radar

Abstract

The application is suitable for the technical field of parking management, and provides an entrance and exit anti-blocking vehicle management system based on a camera and a laser radar, a barrier gate, at least one vehicle identification camera and a laser radar, the camera and the laser radar are arranged at a certain distance from the barrier gate. Specifically, the license plate of the vehicle driving to the barrier gate is pre-identified and judged for legality by the camera, and a license plate queue is established, and the vehicle behavior is tracked and positioned based on the three-dimensional laser radar, the legality of each license plate in the license plate queue is marked, and finally the vehicle barrier gate access permission control is triggered, and the continuous legal vehicle passes through the barrier gate at a time; therefore, the intelligent fusion of the three-dimensional laser radar and the camera is realized, the intelligent judgment of the queuing of multiple vehicles at the peak time is realized, the multiple legal vehicles pass through the barrier gate at a time, the congestion of the peak entrance and exit is avoided, the frequency of lifting and lowering the barrier gate is reduced, and the service life of the equipment is increased.

Description

Technical Field

[0001] This invention belongs to the field of parking management technology, and in particular relates to an entrance and exit anti-congestion vehicle management system based on cameras and lidar. Background Technology

[0002] Currently, vehicle management systems are developing towards greater intelligence. Typical vehicle management systems rely on inductive loop detectors to trigger a method (card swipe, license plate recognition, etc.) to collect vehicle information, then verify the vehicle and control access. Once authorized, the system automatically raises or lowers the barrier. This barrier control method has indeed solved many problems and facilitated the management of many parking areas.

[0003] With the development of technology, new technologies are being applied more and more in vehicle management systems. In particular, the development of artificial intelligence (AI) technology, including object detection, object recognition, object tracking, trajectory calculation, and behavior understanding, is gaining momentum. Among these, object detection and recognition are the foundation for subsequent tracking, trajectory calculation, and behavior understanding.

[0004] 3D lidar is a radar system that uses laser beams to detect the three-dimensional spherical coordinates of targets. Employing coherent angle measurement and frequency-modulated ranging principles, it transmits a detection signal to the target, then compares the received echo signal reflected back from the target with the transmitted signal. After appropriate processing, the target's spherical coordinates can be obtained. With the rapid development of 3D imaging technology, lidar is being applied in various fields. In real-world environments, entrances / exits and vehicle behavior exhibit certain motion trajectories and spatial information. Therefore, introducing laser sensors to acquire the target's three-dimensional information within the scene enables more accurate tracking, trajectory calculation, and behavior understanding.

[0005] Although current parking lot vehicle access management systems are quite mature, most of them are based on camera-based license plate recognition. When a vehicle approaches the rising barrier, the camera captures a photo of the front of the vehicle and sends it back to the system platform. The system platform automatically identifies the license plate number and determines whether the vehicle is legitimate, such as whether it has paid the fee or is an authorized vehicle. If the vehicle is legitimate, the barrier is raised. After the vehicle has completely passed, the barrier is lowered to wait for the next vehicle.

[0006] A common scenario is queuing at entrances and exits. After a vehicle is successfully recognized and completely driven away, sometimes the barrier needs to be fully lowered before the next vehicle can be recognized. Some vehicle management systems, however, simultaneously recognize the license plate of the next vehicle while the barrier is lowering; if the vehicle is legitimate, it raises again. Regardless of the method, license plate recognition proceeds only after the preceding vehicle has passed through. However, license plate recognition is time-consuming and can easily cause congestion at intersections. Furthermore, the sudden raising of the barrier during its descent may accidentally hit vehicles behind it and shorten the lifespan of the barrier.

[0007] Therefore, the functions of cameras and lidar can be used to intelligently judge the queuing of multiple vehicles during peak hours, avoid congestion at entrances and exits during peak hours, and at the same time extend the service life of the equipment and reduce maintenance costs. Summary of the Invention

[0008] In view of the above problems, the purpose of this invention is to provide an entrance and exit anti-congestion vehicle management system based on cameras and lidar, which aims to solve the technical problems of easy congestion at entrances and exits and short service life of existing vehicle management systems during peak hours.

[0009] The present invention adopts the following technical solution:

[0010] The entrance / exit anti-congestion vehicle management system based on cameras and lidar includes a barrier gate, at least one vehicle recognition camera, and lidar. The camera and lidar are positioned at a certain distance from the barrier gate. The implementation method of the vehicle management system is as follows:

[0011] Step S1: Calibrate the parameters of the camera and the lidar, perform deep fusion of lidar point cloud data and camera image data, complete the spatial alignment of lidar and camera, and realize the association between lidar point cloud data and camera image data.

[0012] Step S2: For vehicles entering the field of view, the camera performs license plate recognition, and the LiDAR tracks the vehicle's trajectory. The license plate recognition determines whether the vehicle is legal, and the license plate number and the corresponding legal marker are added to the license plate queue in sequence. The license plate queue is dynamically updated. The legal marker can be either legal or illegal.

[0013] Step S3: Through trajectory tracking, when a vehicle's coordinates are located at the trigger position of the barrier gate, determine the current vehicle's valid mark in the license plate queue;

[0014] Step S4: If the vehicle is an illegal vehicle, the barrier gate will not activate and a prompt will be given.

[0015] Step S5: If the vehicle is legitimate, the barrier gate is raised, waiting for the current consecutive number of legitimate vehicles in the license plate queue to pass through the barrier gate at once. After all vehicles have passed through, the barrier gate is lowered. For vehicles that have already left the barrier gate, they are removed from the license plate queue and tracking is stopped.

[0016] Furthermore, the implementation method of the vehicle management system also includes:

[0017] Step S6: For illegal vehicles, after the business processing is completed, the barrier gate is raised. After the current vehicle passes through the barrier gate, if the next vehicle in the license plate queue is a legal vehicle, the barrier gate remains raised, waiting for the current consecutive number of legal vehicles in the license plate queue to pass through the barrier gate at once; otherwise, the barrier gate is lowered.

[0018] Furthermore, each license plate data in the license plate queue includes the license plate number and a legality mark. During the real-time trajectory tracking of vehicles, if the lidar detects that the distance between two adjacent legal vehicles in the license plate queue is greater than a preset threshold, an interruption mark is added to the license plate data of the following vehicle in the license plate queue. While waiting for a continuous number of legal vehicles to pass through the barrier gate, if the next vehicle is an illegal vehicle, or has an interruption mark, or has not reached the trigger position of the barrier gate within a certain period of time, the barrier gate will then descend.

[0019] Furthermore, in step S2, a default recognition area is set for the barrier gate. Vehicles whose license plates are recognized are first added to a temporary queue. When a vehicle is tracked and enters the default recognition area, the vehicle is inserted from the temporary queue into the license plate queue according to the distance between the vehicle and the barrier gate, and the license plate queue is dynamically updated.

[0020] Furthermore, the default recognition area is rectangular in shape. The average offset distance between the vehicles in the default recognition area and the central axis of the rectangle is calculated. If the average offset distance is greater than the offset threshold, the convoy direction is determined based on the vehicle coordinates. The coordinates of the vehicles in the convoy direction are connected to form a broken line segment. Using the broken line segment as the center line, a detection area of ​​a certain width is generated on both sides. This detection area is the adjusted default recognition area.

[0021] Furthermore, the distance between two vehicles can be determined by the straight-line distance between the coordinates of two adjacent legal vehicles in the license plate queue.

[0022] Furthermore, when a vehicle enters the default recognition area from the left or right side of the temporary queue:

[0023] If there are no vehicles in the default recognition area, the inserted vehicle will be directly transferred to the license plate queue.

[0024] If there is more than one vehicle in the default recognition area, find the vehicle that is closest to the inserted vehicle and record it as the approaching vehicle. The distance between them is recorded as the approaching distance. If the approaching distance is greater than or equal to the set value, the inserted vehicle will be transferred to the license plate queue according to the distance between the inserted vehicle and the barrier gate.

[0025] If the approach distance and the distance to the inserted vehicle are less than the set value, take a fixed distance in front of the approaching vehicle as the starting point, calculate the distance between the starting point and the inserted vehicle and the approaching vehicle, and transfer the inserted vehicle to the vehicle queue according to the distance.

[0026] Furthermore, when a vehicle leaves the default recognition area in the license plate queue, the license plate data of the corresponding vehicle is removed from the license plate queue.

[0027] The beneficial effects of this invention are as follows: This invention utilizes the fusion of video data from LiDAR and cameras to achieve vehicle access control for multiple vehicles near adjacent entrances and exits. Specifically, the camera pre-captures and identifies the license plates of vehicles approaching the barrier gate, determining their legality and establishing a license plate queue. Simultaneously, based on 3D LiDAR, vehicle behavior is tracked and located, and the legality of each license plate in the queue is marked. Finally, in the vehicle access control at the barrier gate, consecutive legal vehicles are triggered to pass through the barrier gate at once. Therefore, this invention, through the intelligent fusion of 3D LiDAR and cameras, achieves intelligent judgment of multiple vehicles queuing during peak hours, triggering multiple legal vehicles to pass through the barrier gate at once, avoiding congestion at entrances and exits during peak hours, while also reducing the frequency of barrier raising and lowering, and increasing the service life of the equipment. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the layout of the vehicle management system provided in an embodiment of the present invention;

[0029] Figure 2 This is a flowchart of the implementation method of the vehicle management system provided in the embodiments of the present invention;

[0030] Figure 3 This is a diagram illustrating the adjustment of the default recognition area;

[0031] Figure 4 This is a schematic diagram illustrating the calculation method for determining the vehicle's position.

[0032] Figure 5 This is a schematic diagram of the second method for determining the vehicle position. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0034] To illustrate the technical solution described in this invention, specific embodiments are described below.

[0035] like Figure 1 As shown, the entrance / exit anti-congestion vehicle management system based on cameras and LiDAR provided in this embodiment includes a barrier gate 100, at least one vehicle recognition camera 200, and a LiDAR 300. In this embodiment, the number of cameras and LiDARs is not limited; the number determines the ability to simultaneously identify and track the number of vehicles, and can be flexibly set according to cost requirements. The cameras and LiDARs are positioned away from the barrier gate, generally mounted on a pole. The cameras here have different functions from ordinary cameras on the barrier gate. Additionally, there are also ordinary cameras on the barrier gate, used for regular barrier gate operations in case of system recognition or tracking anomalies or malfunctions, including license plate recognition errors.

[0036] Based on the above system, this embodiment provides a method for implementing a vehicle management system, such as... Figure 2 As shown, it includes the following steps:

[0037] Step S1: Calibrate the parameters of the camera and LiDAR, perform deep fusion of LiDAR point cloud data and camera image data, complete the spatial alignment of LiDAR and camera, and realize the association between LiDAR point cloud data and camera image data.

[0038] For the installed cameras and LiDAR, the first step is to calibrate their parameters to establish a correlation between the LiDAR point cloud data and the camera image data. Calibration parameters include the camera's intrinsic and extrinsic parameters, as well as the extrinsic parameters between the camera and the LiDAR.

[0039] First, determine the camera's focal length, principal point, and nonlinear distortion coefficients. Then, use the calibrated nonlinear distortion coefficients for image correction to generate image data with a normal viewpoint. A checkerboard pattern is typically used to establish a reference coordinate system. Using the corrected image data, calculate the camera's position relative to the checkerboard pattern to determine its relative position. Similarly, a checkerboard pattern is used as a calibration board between the camera and the LiDAR. Calculate the relative position of the checkerboard pattern in the camera's visual imaging to determine the coordinates of the plane containing the checkerboard pattern in the camera's coordinate system. These coordinates can be represented by the plane's unit normal vector and distance.

[0040] Then, the image data from the camera and the point cloud data from the LiDAR are fused and labeled to obtain the transformation relationship between the coordinate system of the LiDAR and the two-dimensional plane coordinate system of the camera, and the data is labeled using a labeling tool.

[0041] This step establishes a reasonable LiDAR coordinate system and a camera coordinate system. By utilizing the spatial constraints between the LiDAR scanning points and the camera image, the spatial transformation relationship between the two coordinate systems can be solved, thereby achieving spatial alignment between the LiDAR and the camera. This enables the association between LiDAR data and the camera's visible light image, completing the binding of license plate recognition and vehicle tracking. Since the fusion technology of cameras and LiDAR is relatively mature, it will not be elaborated upon here.

[0042] Step S2: For vehicles entering the field of view, the camera performs license plate recognition, and the LiDAR tracks the vehicle's trajectory. The license plate recognition determines whether the vehicle is legal, and the license plate number and corresponding legal marker are added to the license plate queue in sequence. The license plate queue is dynamically updated. The legal marker can be either legal or illegal.

[0043] This embodiment first requires establishing a license plate queue to store vehicle information, including at least the license plate number and a legality marker. Since cameras and LiDAR have a limited range for vehicle identification, only vehicles entering that range will be identified and tracked. Legal vehicles are defined as those that can be allowed to pass directly, while illegal vehicles are defined as those requiring processing before release, such as payment or visitor registration. In a specific application example, within a factory area, vehicles registered inside the factory do not need to pay, while vehicles outside the factory need to pay according to a set time. During the vehicle's passage through the gate, the camera captures a photo of the vehicle's license plate for identification. After obtaining the license plate number, it is compared with the backend data to determine the vehicle's legality. If the vehicle is legal, it is marked as legal; otherwise, it is marked as illegal.

[0044] While the camera captures and identifies the license plate, the LiDAR tracks the vehicle's trajectory. The LiDAR can track multiple vehicles simultaneously, and the coordinates can be mapped onto a plane.

[0045] In one approach, for each identified license plate number, the license plate number and its corresponding valid identifier are entered into the license plate queue in order of distance from the barrier gate. The license plate queue stores multiple license plate records, each including the license plate number and valid identifier. The license plate queue is dynamically updated, including newly added vehicles, vehicles making U-turns before exiting the barrier gate, and adjustments to the order of vehicles in the queue.

[0046] Step S3: Through trajectory tracking, when a vehicle's coordinates are located at the trigger position of the barrier gate, determine the current vehicle's legal status in the license plate queue.

[0047] Assuming the barrier gate is currently in the lowered state, during tracking, if a vehicle is located at the gate's entrance / exit position, its license plate number in the license plate queue is checked for legitimacy. Trigger coils can typically be installed at the gate's entrance / exit positions to automatically detect vehicle entry.

[0048] Step S4: If the vehicle is an illegal vehicle, the barrier gate will not move and a prompt will be given.

[0049] Step S5: If the vehicle is legitimate, the barrier gate is raised, waiting for the current consecutive number of legitimate vehicles in the license plate queue to pass through the barrier gate at once. After all vehicles have passed through, the barrier gate is lowered. For vehicles that have already left the barrier gate, they are removed from the license plate queue and tracking is stopped.

[0050] If the vehicle is legitimate, the barrier gate will automatically lift. The system will pre-read the legitimacy markers of subsequent vehicles in the license plate queue. For example, if three consecutive vehicles are legitimate and the fourth is illegitimate, the barrier gate will remain raised and wait for the three legitimate vehicles to pass through at once before automatically lowering it. When the fourth vehicle reaches the trigger position, appropriate processing can be performed. After processing, the barrier gate will lift. After the current vehicle passes through, the system will continue to check the license plate data in the queue to determine whether the barrier gate should lower. If the next vehicle in the queue is legitimate, the barrier gate will remain raised and wait again for the current consecutive number of legitimate vehicles in the queue to pass through at once; otherwise, the barrier gate will lower. Throughout this process, vehicles that leave the barrier gate will be removed from the license plate queue and tracking will cease.

[0051] Therefore, by using the above methods, the number of times the barrier gate is raised and lowered can be significantly reduced, thereby improving traffic flow. In particular, for application scenarios where there are long queues waiting to enter or exit, vehicles can pass through quickly, avoiding congestion at entrances and exits.

[0052] Furthermore, in practical applications, to avoid situations where the distance between two adjacent legitimate vehicles is too great, or a legitimate vehicle behind is traveling too slowly, causing the barrier to remain raised for an extended period and potentially leading to parking lot security issues, a preferred approach is to use a system where each license plate in the license plate queue includes the license plate number and a legitimacy marker. During real-time vehicle tracking, the LiDAR can detect the distance between two adjacent legitimate vehicles in the queue; here, the straight-line distance between the coordinates of the two legitimate vehicles represents their actual distance. If the distance exceeds a preset threshold, an interruption marker is added to the license plate data of the following vehicle in the queue. While waiting for a continuous number of legitimate vehicles to pass through the barrier, if the next vehicle after the current vehicle passes is an illegitimate vehicle, has an interruption marker, or fails to reach the barrier's trigger position within a certain period, the barrier will immediately lower. This method effectively solves the problem of the barrier remaining raised for an extended period due to excessive distance between two vehicles or slow vehicle speed.

[0053] If the camera and lidar are close to the barrier gate, the number of vehicles that can be identified and tracked is limited. To increase the number of vehicles that can be identified and tracked, more cameras and lidar can be set up or placed at slightly greater distances. It is also unnecessary to add all identified vehicles to the license plate queue, because in reality, some vehicles may cut in from the side, or turn around immediately before entering or exiting the barrier gate. Vehicles at the end of the queue (i.e., farther from the barrier gate) often enter the queue side-by-side, causing their positions to fluctuate greatly and making it difficult to accurately determine their order. Therefore, it is unnecessary to add all these vehicles to the license plate queue in order of distance. Furthermore, since the license plate queue is dynamically updated, to improve management efficiency, this embodiment further establishes a temporary queue and a default recognition area for the barrier gate, such as... Figure 1 The area shown is 400. The default recognition area is generally rectangular. Only vehicles entering the default recognition area are added to the license plate queue. Vehicles in the temporary queue are not distinguished by their front or back positions.

[0054] Therefore, when a vehicle's license plate is identified, it is first added to a temporary queue. Once the vehicle enters the default recognition area, it is then inserted into the license plate queue according to its distance from the barrier gate. This allows for dynamic updates to the license plate queue. This is because vehicles in the temporary queue may enter the default recognition area from the side or at an angle, making it impossible to simply update the queue based on the order of their entry into the default recognition area. Instead, it is sufficient to sort the vehicles entering the default recognition area by their positions, reducing update complexity and enabling faster response.

[0055] Under normal circumstances, since vehicles are usually arranged in a straight line, the default recognition area is rectangular. However, in some special cases, obstacles may obstruct the queue, causing it to deviate from a straight line. For example... Figure 3 As shown, in this scenario, the shape of the default recognition area needs to be adjusted to accurately identify vehicles entering it. Specifically, the average offset distance between vehicles within the default recognition area and the rectangular centerline is calculated. If the average offset distance is greater than an offset threshold, the queuing direction is determined based on the vehicle coordinates. A broken line segment is formed by connecting the coordinates of vehicles in the queuing direction. Using this broken line segment as the centerline, a detection area of ​​a certain width is generated on both sides; this detection area is the adjusted default recognition area. In a typical straight-line queue, the deviation between a vehicle and the rectangular centerline can be positive or negative. Even if there is a deviation in the queue position, the overall average offset distance from the centerline is relatively small. If the average offset distance significantly increases, the default recognition area needs to be adjusted according to the queueing direction.

[0056] Another issue is vehicle queue jumping. Vehicles outside the default recognition zone may cut into the queue from either side or one side. Setting a default recognition zone can detect queue jumping dynamics, including approximating whether queue jumping is possible. Normally, for vehicles inserting into a temporary queue, if they are outside the default recognition zone, it is considered that the inserting vehicle cannot currently complete the queue jump. When they enter the default recognition zone, it means that the inserting vehicle is close enough to the queued vehicles to complete the queue jump.

[0057] Specifically, there are two situations:

[0058] 1. If there are no vehicles in the default recognition area, the inserted vehicle will be directly transferred to the license plate queue. That is, the inserted vehicle will be deleted from the temporary queue and added to the license plate queue.

[0059] 2. If there is more than one vehicle in the default recognition area, find the vehicle closest to the inserted vehicle and mark it as the approaching vehicle. The distance between them is recorded as the approach distance. If the approach distance is greater than or equal to a set value, the inserted vehicle is transferred to the license plate queue based on the distance between the inserted vehicle and the barrier gate. If the approach distance is less than the distance between the inserted vehicle and the approaching vehicle, take a fixed distance in front of the approaching vehicle as the starting point, calculate the distance between the starting point and the inserted vehicle and the approaching vehicle, and transfer the inserted vehicle to the vehicle queue according to the distance.

[0060] When a vehicle enters the default recognition area, the system can locate the nearest vehicle based on its coordinates; this is called the approaching vehicle, and the distance is called the approach distance. If the approach distance is greater than or equal to a set value, it means the vehicle is too far away, and the vehicle can be directly moved to the license plate queue based on its distance from the barrier gate.

[0061] If the approach distance and the distance to the inserted vehicle are less than the set value, it means the vehicles are very close. Figure 4 As shown (hollow circles represent approaching vehicles, solid circles represent vehicles that have entered the gate), it is difficult to determine the relative positions of the two vehicles based solely on their distance from the gate, because the gate is relatively far away while the two vehicles are approximately close to it. In this situation, if... Figure 5 As shown, a fixed distance in front of the vehicle is set as the starting point. Without using the barrier gate as a reference, the distance between the inserted vehicle and the starting point is compared with the distance between the approaching vehicle and the starting point to ensure the inserted vehicle is positioned appropriately. A significant advantage of this method is that moving the starting point forward facilitates comparison of distance differences. Figure 5 As shown, the distance between the inserting vehicle and the starting point must be greater than that of the approaching vehicle; therefore, the approaching vehicle needs to be positioned behind it. This method also has another advantage: in actual queue-jumping, vehicles traveling on the main line have an advantage, such as... Figure 5As shown in the diagram, the vertical distance between the inserting vehicle and the barrier gate is less than that of the approaching vehicle. Theoretically, the inserting vehicle should be in front. However, considering the actual situation of queue jumping, even if the inserting vehicle is slightly ahead of the approaching vehicle, it will not actually be inserted in front of the approaching vehicle. Therefore, this method is simple to determine, effective, and close to the actual situation.

[0062] Finally, when a vehicle leaves the default recognition area in the license plate queue, the license plate data of the corresponding vehicle is removed from the license plate queue, thus updating the data in the license plate queue.

[0063] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An entrance / exit anti-congestion vehicle management system based on cameras and lidar, characterized in that, The vehicle management system includes a barrier gate, at least one vehicle recognition camera, and a lidar sensor. The camera and lidar sensor are positioned at a certain distance from the barrier gate. The implementation method of the vehicle management system is as follows: Step S1: Calibrate the parameters of the camera and the LiDAR, perform deep fusion of the LiDAR point cloud data and the camera image data, complete the spatial alignment of the LiDAR and the camera, and realize the association between the LiDAR point cloud data and the camera image data. Step S2: For vehicles entering the field of view, the camera performs license plate recognition, and the LiDAR tracks the vehicle's trajectory. The license plate recognition determines whether the vehicle is legal, and the license plate number and the corresponding legal marker are added to the license plate queue in sequence. The license plate queue is dynamically updated. The legal marker can be either legal or illegal. Step S3: Through trajectory tracking, when a vehicle's coordinates are located at the trigger position of the barrier gate, determine the current vehicle's valid mark in the license plate queue; Step S4: If the vehicle is an illegal vehicle, the barrier gate will not activate and a prompt will be given. Step S5: If the vehicle is legitimate, the barrier gate is raised, and the gate is opened to allow the current number of legitimate vehicles in the license plate queue to pass through at once. Once all vehicles have passed through, the barrier gate is lowered. Vehicles that have already exited the gate are removed from the license plate queue and tracking is stopped. Each license plate data in the license plate queue includes the license plate number and a legality mark. During the real-time trajectory tracking of vehicles, if the laser radar detects that the distance between two adjacent legal vehicles in the license plate queue is greater than a preset threshold, an interruption mark is added to the license plate data of the following vehicle in the license plate queue. While waiting for a continuous number of legal vehicles to pass through the barrier gate, if the next vehicle is an illegal vehicle, or has an interruption mark, or has not reached the trigger position of the barrier gate within a certain period of time, the barrier gate will then descend. In step S2, a default recognition area is set for the barrier gate. Vehicles whose license plates are recognized are first added to a temporary queue. When a vehicle is tracked and enters the default recognition area, the vehicle is inserted from the temporary queue into the license plate queue according to the distance between the vehicle and the barrier gate, and the license plate queue is dynamically updated. The default recognition area is rectangular in shape. The average offset distance between the vehicles in the default recognition area and the central axis of the rectangle is calculated. If the average offset distance is greater than the offset threshold, the direction of the convoy is determined based on the vehicle coordinates. The coordinates of the vehicles in the convoy direction are connected to form a broken line segment. The broken line segment is used as the center line to generate a detection area of ​​a certain width on both sides. This detection area is the adjusted default recognition area. When a vehicle enters the default recognition area from the left or right side of the temporary queue: If there are no vehicles in the default recognition area, the inserted vehicle will be directly transferred to the license plate queue. If there is more than one vehicle in the default recognition area, find the vehicle that is closest to the inserted vehicle and record it as the approaching vehicle. The distance between them is recorded as the approaching distance. If the approaching distance is greater than or equal to the set value, the inserted vehicle will be transferred to the license plate queue according to the distance between the inserted vehicle and the barrier gate. If the approach distance and the distance to the inserted vehicle are less than the set value, take a fixed distance in front of the approaching vehicle as the starting point, calculate the distance between the starting point and the inserted vehicle and the approaching vehicle, and transfer the inserted vehicle to the vehicle queue according to the distance.

2. The entrance / exit anti-congestion vehicle management system based on cameras and lidar as described in claim 1, characterized in that, The implementation method of the vehicle management system further includes: Step S6: For illegal vehicles, after the business processing is completed, the barrier gate is raised. After the current vehicle passes through the barrier gate, if the next vehicle in the license plate queue is a legal vehicle, the barrier gate remains raised, waiting for the current consecutive number of legal vehicles in the license plate queue to pass through the barrier gate at once; otherwise, the barrier gate is lowered.

3. The entrance / exit anti-congestion vehicle management system based on cameras and lidar as described in claim 1, characterized in that, The distance between two vehicles is determined by the straight-line distance between the coordinates of two adjacent legal vehicles in the license plate queue.

4. The entrance / exit anti-congestion vehicle management system based on cameras and lidar as described in claim 1, characterized in that, When a vehicle leaves the default recognition area in the license plate queue, the license plate data of the corresponding vehicle is removed from the license plate queue.

Images