A Tunnel Vehicle-Assisted Localization Method Based on Roadside Video Stream Trajectory Tracking

Abstract

This invention relates to a tunnel vehicle-assisted positioning method based on roadside video stream trajectory tracking, comprising the following steps: S1, setting up a high-definition license plate recognition system at the tunnel entrance section, setting up a high-definition camera in an overlapping monitoring area, and equipping the tunnel roadside with a Roadside Unit (RSU); S2, completing the tracking of the target vehicle and recording the temporary sequential trajectory ID of the formed trajectory; S3, performing license plate recognition on vehicles in the license plate recognition area captured by the high-definition license plate recognition system; S4, converting all vehicle horizontal and vertical coordinates to Frenet coordinates, matching the video trajectory with the license plate ID in the same coordinate system, and replacing the temporary sequential trajectory ID with the license plate ID; S5, when a vehicle enters the multi-camera overlapping monitoring area, completing the handover of monitoring between the first and second cameras; S6, using the traffic information transceiver module of the RSU unit to broadcast information to surrounding vehicles according to frequency; S7, vehicles equipped with an On-Board Unit (OBU) receive the information broadcast by the RSU. This invention can solve the problem of navigation systems failing to function properly when vehicle satellite positioning fails in tunnels.

Description

Technical Field

[0001] This invention relates to the field of smart highways and vehicle-road cooperative applications, and more specifically, to a tunnel vehicle-assisted positioning method based on roadside video stream trajectory tracking. Background Technology

[0002] Roadside video stream tracking refers to the continuous deployment of video cameras at regular intervals along one or both sides of a road to monitor road conditions. Currently, considerable research utilizes roadside equipment as a technological means to identify and track target vehicles, thereby obtaining vehicle location information. This includes microwave radar, lidar, video holographic sensing devices, and RFID electronic tags. Microwave radar devices have a fan-shaped acquisition range but often collect interference information; lidar is more expensive than microwave radar; and RFID electronic tags can only provide location information for fixed points. Therefore, compared to other roadside equipment, video camera equipment, because it better reflects the real conditions of the road surface and vehicles, is often used alone or in combination with other roadside equipment to form holographic sensors for traffic condition monitoring.

[0003] Real-time vehicle location services provided by smartphones and pre-installed vehicle navigation software have become an important means for drivers to obtain real-time route information. However, providing vehicle location services in tunnels faces the following challenges: 1) It requires existing smartphones as terminals and existing navigation software as a foundation; 2) Global Navigation Satellite System (GNSS) signals are lost in tunnels, rendering navigation software unusable; 3) Vehicles travel at higher speeds in tunnels, potentially leading to signal loss, reduced positioning accuracy, and increased latency. Therefore, it is necessary to study tunnel-assisted positioning technology from the perspective of vehicle-road cooperation to provide basic positioning services for travelers and to assist in intelligent monitoring of traffic conditions and safe driving guidance within tunnels. This also represents a crucial requirement for informatization, digitalization, and safe driving in the field of smart highway construction.

[0004] However, existing roadside technologies mainly focus on vehicle positioning research within a certain cross section or the monitoring range of a single device, with relatively few studies on continuous positioning-assisted navigation for vehicles in long tunnels where global navigation satellite systems fail. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a vehicle-assisted positioning method in tunnels based on roadside video stream trajectory tracking, which can solve the problem of vehicle satellite positioning failure and navigation system malfunction in tunnels, and provide a safe, stable, lane-level accurate vehicle-assisted positioning method by relying on the road.

[0006] The technical solution adopted by this invention to solve its technical problem is: to construct a tunnel vehicle-assisted positioning method based on roadside video stream trajectory tracking, comprising the following steps:

[0007] S1. A high-definition license plate recognition system is installed at the tunnel entrance section, and high-definition cameras are installed on the tunnel roadside at intervals of 200-250m. Adjacent high-definition cameras are set up with an overlapping monitoring area of ​​25-35m, and RSU units are equipped on the tunnel roadside.

[0008] S2. Use video detection algorithm to detect vehicles in the video captured by the high-definition camera and complete the tracking of the target vehicle. Extract the horizontal coordinate, vertical coordinate, timestamp, driving direction, speed, vehicle length and lane information of each frame of the tracked target vehicle, and record the temporary pipeline trajectory ID that forms the trajectory.

[0009] S3. Perform license plate recognition on vehicles in the license plate recognition area captured by the high-definition license plate recognition system, and extract license plate ID, speed, vehicle length, driving direction, lane, timestamp, horizontal coordinate and vertical coordinate from the recognized vehicles;

[0010] S4. In the tunnel roadside calculation unit, a unified coordinate system is established, and all vehicle horizontal and vertical coordinates are converted to Frenet coordinates. Video trajectories are matched with license plate IDs under the same coordinate system, and the temporary trajectory serial ID is replaced with the license plate ID.

[0011] S5. When the vehicle enters the multi-camera overlapping monitoring area, the second camera trajectory is spliced ​​with the first camera trajectory in the tunnel roadside calculation unit to complete the handover of the first and second camera monitoring. The control variables in the splicing are not limited to timestamps, geographical coordinates, driving direction and lane.

[0012] S6. Using the traffic information transceiver module of the RSU unit, broadcast information to surrounding vehicles at a certain frequency. Each message includes key information such as timestamp, license plate ID, coordinates of the corresponding ID vehicle, vehicle length, driving direction and speed.

[0013] S7. Vehicles equipped with OBU receive information broadcast by RSU and parse the real-time location information belonging to the vehicle based on the license plate ID for in-vehicle navigation.

[0014] According to the above scheme, in step S1, the spacing of the high-definition cameras is not limited and is adjusted according to the actual monitoring range of the equipment. The high-definition cameras are set on the tunnel sidewall or tunnel ceiling, and the license plate recognition system is set at the tunnel entrance. Adjacent high-definition cameras are set with a 25-35m overlapping monitoring area. The length of the overlapping area meets the minimum requirement of 1.5 seconds for multi-camera trajectory stitching. The RSU unit on the tunnel roadside is equipped with a traffic information transceiver module and a calculation module.

[0015] According to the above scheme, in step S2, the algorithm for vehicle detection in the video captured by the high-definition camera is not limited, and the algorithm for vehicle target tracking is not limited; the extracted traffic variable data includes key information such as the target timestamp, license plate ID, horizontal and vertical coordinates of the corresponding ID vehicle, vehicle length, driving direction and speed for each frame, and forms a temporary ID for video trajectory recording.

[0016] According to the above scheme, in step S3, there are no limitations on the high-definition license plate recognition system equipment, which can adapt to various complex weather conditions; the high-definition license plate recognition system extracts information such as license plate ID, speed, vehicle length, driving direction, lane, timestamp, horizontal coordinate and vertical coordinate.

[0017] According to the above scheme, in step S4, the vehicle positions under the high-definition camera plane coordinate system and the license plate recognition system plane coordinate system are all converted into Frenet coordinates based on the current road alignment; under the Frenet coordinate system, the video trajectory of the nearest vehicle at the same time is matched according to the license plate ID detection timestamp. The trajectory meets the conditions of having the same driving lane, driving direction, and vehicle length information; otherwise, it is considered an invalid match; for the successfully matched trajectory, the trajectory temporary serial ID is replaced with the license plate ID, and the coordinates are converted into latitude and longitude coordinates. This process is completed by the roadside calculation unit.

[0018] According to the above scheme, in step S5, sufficient overlap area is retained in the multi-camera overlapping monitoring. The roadside calculation unit recognizes the same trajectory based on the centroid coordinates of vehicles at different camera positions under the same timestamp within a continuous frame if the distance meets the set threshold. The trajectory meets the condition of having the same driving lane, driving direction, and vehicle length information. The handover of monitoring between the first camera position and the second camera position is completed, and the license plate ID information of the previous camera position is inherited.

[0019] According to the above scheme, in step S6, the trajectory information calculated and associated by the roadside computing unit includes key information such as timestamp, license plate ID, latitude and longitude coordinates of the corresponding ID vehicle, vehicle length, driving direction, and speed. This information is then broadcast to surrounding vehicles through the traffic information transceiver module of the RSU unit in accordance with the device communication protocol.

[0020] According to the above scheme, in step S7, the vehicle equipped with OBU receives information broadcast from RSU, and the license plate ID is used as a unique identification information for data docking and matching. The vehicle OBU only extracts the real-time location information of the vehicle itself for vehicle navigation when the vehicle GPS signal is weakened.

[0021] The tunnel vehicle-assisted positioning method based on roadside video stream trajectory tracking of the present invention has the following beneficial effects:

[0022] 1. This invention uses multiple video surveillance cameras continuously deployed along the roadside to detect the location information of passing vehicles, and uses a license plate recognition system at the tunnel entrance to confirm the vehicle's identity information. A computing unit associates the driving trajectory with the license plate, and the RSU unit and on-board OBU broadcast the vehicle's geographical coordinates to the vehicle to assist navigation in tunnels or other scenarios where the GPS signal is weakened. This method of comprehensively utilizing roadside video surveillance and on-board communication equipment can not only make up for the shortcomings of the GPS signal in complex scenarios such as tunnels, but also provide more accurate and reliable vehicle positioning information, thereby greatly improving the safety and reliability of vehicle driving.

[0023] 2. The multi-camera video surveillance camera of this invention also has other functions, including but not limited to road condition monitoring, traffic incident alarms, event playback, and traffic flow parameter extraction. These functions make the construction of smart highways more comprehensive and diversified, not only helping to monitor traffic conditions and improve traffic management efficiency, but also providing important data support for road maintenance and accident handling; therefore, roadside video surveillance cameras have become one of the necessary equipment for the construction of smart highways, and their application in vehicle-assisted positioning provides important support for the improvement of smart transportation systems.

[0024] 3. The vehicle-assisted positioning method proposed in this invention is not only of great significance in solving the problem of insufficient GPS signal, but also provides an innovative solution for smart highway construction and traffic management, and has a profound impact on improving road traffic safety, smoothness and intelligence. Attached Figure Description

[0025] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings:

[0026] Figure 1 This is a flowchart of the tunnel vehicle assisted positioning method based on roadside video stream trajectory tracking according to the present invention;

[0027] Figure 2 This is a schematic diagram of the deployment of the roadside continuous multi-camera video surveillance equipment of the present invention;

[0028] Figure 3 This is a schematic diagram illustrating the operation of the roadside multi-camera video surveillance-assisted positioning system of the present invention. Detailed Implementation

[0029] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0030] like Figure 1-3 As shown, the tunnel vehicle-assisted positioning method based on roadside video stream trajectory tracking of the present invention includes the following steps:

[0031] S1. Deploy the high-definition license plate recognition system at the tunnel entrance section. Deploy high-definition cameras at certain intervals (200-250m) on the tunnel roadside. Set up an overlapping monitoring area of ​​25-35m between adjacent cameras and equip them with RSU units on the roadside.

[0032] There are no restrictions on the spacing of high-definition cameras; the spacing should be adjusted according to the actual monitoring range of the equipment. High-definition cameras should, in principle, be installed on the tunnel sidewalls or tunnel ceiling. License plate recognition systems should, in principle, be installed near the tunnel entrance, but additional systems can be added inside the tunnel as needed. Adjacent cameras should have a certain length of overlapping monitoring area, and the length of the overlapping area should meet the minimum time requirement for multi-camera trajectory stitching. The roadside RSU unit has traffic information transceiver modules, computing modules, etc.

[0033] S2. Use video detection algorithms to detect vehicles in the video captured by the high-definition camera and complete the tracking of the target vehicles. Extract vehicle information such as horizontal coordinate, vertical coordinate, timestamp, driving direction, speed, vehicle length, and lane for each frame of the tracked target vehicle, and record the temporary pipeline trajectory ID that forms the trajectory.

[0034] The algorithm for vehicle detection in video captured by high-definition cameras is not limited, nor is the algorithm for vehicle target tracking. The extracted traffic variable data shall include key information such as the target timestamp, license plate ID, horizontal and vertical coordinates of the corresponding ID vehicle, vehicle length, driving direction, and speed for each frame, and form a temporary ID for video trajectory recording.

[0035] S3. Perform license plate recognition on vehicles in the license plate recognition area captured by the high-definition license plate recognition system, and extract license plate ID, speed, vehicle length, driving direction, lane, timestamp, horizontal coordinate, and vertical coordinate from the recognized vehicles.

[0036] There are no restrictions on the high-definition license plate recognition system equipment, but it should be able to adapt to various complex weather conditions; the license plate recognition system should at least extract information such as license plate ID, speed, vehicle length, driving direction, lane, timestamp, horizontal coordinate (in the license plate system coordinate system), and vertical coordinate (in the license plate system coordinate system).

[0037] S4. In the roadside calculation unit, a unified coordinate system is established, and all vehicle horizontal and vertical coordinates are converted to Frenet coordinates. Video trajectories are matched with license plate IDs under the same coordinate system, and the temporary trajectory serial ID is replaced with the license plate ID.

[0038] The vehicle positions under the high-definition camera's planar coordinate system and the license plate recognition system's planar coordinate system are all converted into Frenet coordinates based on the current road alignment. Under the Frenet coordinate system, the video trajectory of the nearest vehicle at the same time is matched according to the license plate ID detection timestamp. Trajectories that meet the conditions should further satisfy information such as the same driving lane, driving direction, and vehicle length; otherwise, they are considered invalid matches. For successfully matched trajectories, the trajectory temporary serial ID is replaced with the license plate ID, and the coordinates are converted into latitude and longitude coordinates. This process is completed by the roadside calculation unit.

[0039] S5. When the vehicle enters the multi-camera overlapping monitoring area, the roadside calculation unit splices the trajectory of the second camera with the trajectory of the first camera to complete the handover of monitoring between the first and second cameras. The control variables in the splicing are not limited to timestamps, geographical coordinates, driving direction, lanes, etc.

[0040] In multi-camera overlapping monitoring, since sufficient overlap area is retained, the roadside calculation unit recognizes vehicles from different cameras at the same timestamp as having the same trajectory if the distance between them in consecutive frames meets a set threshold. The trajectory that meets the criteria should further satisfy information such as the same driving lane, driving direction, and vehicle length. At this point, the handover between the first and second camera monitoring is completed, and the license plate ID information of the previous camera will be inherited.

[0041] S6. Use the RSU unit traffic information transceiver module to broadcast information to surrounding vehicles at a certain frequency. Each message should include at least the timestamp, license plate ID, coordinates of the corresponding vehicle ID, vehicle length, direction of travel, speed, and other key information.

[0042] The trajectory information obtained by the roadside computing unit includes key information such as timestamp, license plate ID, latitude and longitude coordinates of the corresponding vehicle ID, vehicle length, driving direction, and speed. This information is then broadcast to surrounding vehicles at a certain frequency through the traffic information transceiver module of the RSU unit.

[0043] S7. Vehicles equipped with an OBU receive information broadcast by the RSU and parse the real-time location information belonging to the vehicle based on the license plate ID for in-vehicle navigation. Vehicles equipped with an OBU receive information broadcast from the RSU, and the license plate ID serves as a unique identification information used for data integration and matching. The onboard OBU only extracts the vehicle's real-time location information for in-vehicle navigation when the onboard GPS signal is weakened.

[0044] Finally, it should be noted that the above implementation examples only demonstrate a method for vehicle-assisted positioning under continuous two-camera setups. It should be understood that this method can be used for continuous multi-camera setups and is not limited to the two-camera condition described in the examples. Furthermore, when the road segment is long enough, to ensure the license plate ID is stably inherited as a unique identification ID across continuous multi-camera setups, the license plate recognition system can be supplemented at intervals as needed. Simultaneously, this method uses the license plate ID as vehicle identification information for the interface between the roadside RSU and the on-board OBU. In addition, other methods such as electronic tag RFID (Radio Frequency Identification) wireless identification technology can also be used for identification purposes. These should be understood as variations in detail and form of this method, and do not depart from the scope defined by the claims of this invention.

[0045] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1.A tunnel vehicle assisted positioning method based on roadside video stream trajectory tracking, characterized in that, The method comprises the following steps: S1, setting a high-definition license plate recognition system at the tunnel entrance section, and setting high-definition cameras at the tunnel side at an interval of 200-250 m, and setting an overlapping monitoring area of 25-35 m between adjacent high-definition cameras, and setting an RSU unit at the tunnel side; S2, using a video detection algorithm to detect vehicles in the video captured by the high-definition cameras, and completing tracking of target vehicles, extracting the horizontal coordinate, vertical coordinate, time stamp, driving direction, speed, vehicle length, and lane information of each frame of the tracked target vehicles, and recording the temporary flow trajectory ID of the trajectory; S3, performing license plate recognition on vehicles in the license plate recognition area captured by the high-definition license plate recognition system, and extracting the license plate ID, speed, vehicle length, driving direction, lane, time stamp, horizontal coordinate, and vertical coordinate of the recognized vehicles; S4, converting all the vehicle horizontal and vertical coordinates into Frenet coordinates in the unified coordinate system of the tunnel side calculation unit, and matching the video trajectory and the license plate ID in the same coordinate, and replacing the temporary flow trajectory ID with the license plate ID; S5, when a vehicle enters the multi-camera overlapping monitoring area, the tunnel side calculation unit splices the second camera trajectory and the first camera trajectory, completes the monitoring handover between the first camera and the second camera, and the control variables in the splicing are not limited to the time stamp, geographical position coordinate, driving direction, and lane; S6, using the RSU unit traffic information transceiver module to broadcast information to surrounding vehicles at a certain frequency, and each piece of information includes the time stamp, license plate ID, coordinate of the corresponding ID vehicle, vehicle length, driving direction, and key information of speed; S7, the OBU vehicle receives the information broadcast by the RSU, and parses the real-time position information belonging to the vehicle according to the license plate ID for vehicle navigation. 2.The tunnel vehicle assisted positioning method based on roadside video stream track tracking according to claim 1, characterized in that, In the step S1, the arrangement interval of the high-definition cameras is not limited, and is adjusted according to the actual equipment monitoring range, the high-definition cameras are arranged on the tunnel side wall or the tunnel roof, the license plate recognition system is arranged at the tunnel entrance, the adjacent high-definition cameras are arranged in an overlapping monitoring area of 25-35 m, the length of the overlapping area meets the minimum requirement time of 1.5 seconds for multi-camera trajectory splicing; the RSU unit matched with the tunnel side is provided with a traffic information transceiver module and a calculation module. 3.The tunnel vehicle assisted positioning method based on roadside video stream trajectory tracking according to claim 1, wherein, In the step S2, the vehicle detection algorithm for the video captured by the high-definition cameras is not limited, and the vehicle target tracking algorithm is not limited; the extracted traffic variable data includes the key information of each frame target time stamp, license plate ID, corresponding ID vehicle horizontal and vertical coordinates, vehicle length, driving direction, and speed, and forms a video trajectory record flow temporary ID. 4.The tunnel vehicle assisted positioning method based on roadside video stream trajectory tracking according to claim 1, wherein, In the step S3, the high-definition license plate recognition system is not limited, and can adapt to various complex weather conditions; the high-definition license plate recognition system extracts the information of license plate ID, speed, vehicle length, driving direction, lane, time stamp, horizontal coordinate, and vertical coordinate. 5.The tunnel vehicle assisted positioning method based on roadside video stream trajectory tracking according to claim 1, wherein, In the step S4, the high-definition camera plane coordinate system and the vehicle position in the license plate recognition system plane coordinate system are all converted into Frenet coordinates based on the current road alignment; in the Frenet coordinate system, the nearest vehicle video track at the same time is matched according to the license plate ID detection timestamp, and the track that meets the conditions satisfies the same driving lane, driving direction and vehicle length information, otherwise it is considered as invalid matching; for the matched track, the track temporary serial ID is replaced with the license plate ID, and the coordinates are converted into latitude and longitude coordinates, and the process is completed by the roadside computing unit. 6.The tunnel vehicle assisted positioning method based on roadside video stream trajectory tracking according to claim 1, wherein, In the step S5, in the multi-position overlapping monitoring, a sufficient overlapping area is reserved, and the roadside computing unit determines that the same track is matched according to the distance between the mass center coordinates of the vehicles in different positions under the same timestamp in the continuous frames, and the distance meets the set threshold; the track that meets the conditions satisfies the same driving lane, driving direction and vehicle length information; the first position and the second position monitoring are completed, and the license plate ID information of the previous position is inherited. 7.The tunnel vehicle assisted positioning method based on roadside video stream trajectory tracking according to claim 1, characterized in that, In the step S6, the track information obtained by the roadside computing unit includes the timestamp, the license plate ID, the latitude and longitude coordinates of the corresponding ID vehicle, the vehicle length, the driving direction and the speed, and the information is broadcasted and published to the surrounding vehicles by the RSU unit traffic information transceiver module according to the device communication protocol. 8.The tunnel vehicle assisted positioning method based on roadside video stream trajectory tracking of claim 1, wherein, In the step S7, the OBU vehicle receives the information broadcast from the RSU, and the license plate ID is used as the unique identity information for data docking and matching; the vehicle-mounted OBU only extracts the real-time position information of the vehicle under the condition that the vehicle-mounted GPS signal is weakened, and the information is used for vehicle navigation.

Images