Sensing method and system for roadside device, and device and storage medium

Abstract

A sensing method and system for a roadside device, and an electronic device and a storage medium. The method (200) comprises: producing topographic base map data of an area where a roadside device (100) is located (201); receiving point cloud data that is acquired by means of at least one Lidar sensor (103, 104) disposed on the roadside device (100) (202); performing data registration on the base map data and the point cloud data that is acquired by means of the at least one Lidar sensor (103, 104), so as to obtain fused point cloud data (203); on the basis of the base map data, filtering out ground point clouds from the fused point cloud data, so as to obtain at least one subset containing a single target (204); and extracting a minimum bounding box from the at least one subset containing a single target, so as to obtain a sensing result in the area (205). In the method, high-precision and high-frequency point cloud data is acquired in real time, such that sensing requirements of open-pit mine autonomous driving systems regarding roads outside the systems and regarding target objects are met, thereby ensuring the safety, stability and operation efficiency of autonomous driving systems.

Description

A perception method, system, device and storage medium for roadside equipment

[0001] This disclosure claims priority to the Chinese patent application filed with the State Intellectual Property Office of China on July 26, 2023, with application number 202310923374.X, and invention name “A perception method, system, device and storage medium for roadside equipment”, the entire contents of which are incorporated by reference into this disclosure. Technical Field

[0002] The embodiments of the present disclosure belong to the field of unmanned driving technology, and specifically relate to a perception method, system equipment, and storage medium for roadside equipment. Background Art

[0003] The application of autonomous driving systems in open-pit mines is currently limited to independently controlled sections of road, where only vehicles within the system are allowed to travel. As the range of autonomous vehicles in mining areas expands, intersections have emerged where independently operated sections intersect with sections outside the autonomous driving system, leading to the need for mixed traffic between autonomous and manually driven vehicles at these intersections and sections.

[0004] Due to the complex roads and rugged terrain in mining areas, relying solely on the perception equipment on autonomous vehicles cannot guarantee effective perception of vehicles and other targets outside the autonomous driving system and outside the independently operated sections. Summary of the Invention

[0005] The embodiments of the present disclosure aim to solve at least one of the technical problems existing in the prior art and provide a perception method, system, device and storage medium for roadside equipment.

[0006] In a first aspect of the present disclosure, a method for sensing a roadside device is provided, comprising:

[0007] Creating topographic base map data of the area where the roadside equipment is located;

[0008] Receiving point cloud data acquired by at least one laser radar sensor provided on the roadside equipment;

[0009] Performing data registration on the base map data and the point cloud data acquired by the at least one lidar sensor to obtain fused point cloud data;

[0010] Filtering the ground point cloud from the fused point cloud data based on the base map data, and then segmenting the data after filtering the ground point cloud to obtain at least one subset containing a single target;

[0011] A minimum bounding box is extracted for the at least one subset containing a single target to obtain a perception result within the area.

[0012] Furthermore, the at least one lidar sensor includes a main lidar sensor for acquiring first point cloud data and a blind-complement lidar sensor for acquiring second point cloud data; and the data registration of the base map data with the point cloud data acquired by the at least one lidar sensor includes data fusion of the first point cloud data and the second point cloud data.

[0013] Furthermore, the detection distance or field of view angle of the main lidar sensor and the blind spot compensation lidar sensor are different.

[0014] Optionally, the scanning line beam of the main lidar sensor is larger than the scanning line beam of the blind spot lidar sensor, and the detection range of the main lidar is larger than the detection range of the blind spot lidar sensor.

[0015] Optionally, the scanning beam of the main lidar sensor is more than 64 lines, the detection distance is more than 150 meters, and the scanning angle is less than 120°; the scanning beam of the blind spot lidar sensor is less than 32 lines, the detection distance is less than 50 meters, and the scanning angle is 360°.

[0016] Furthermore, the data registration includes calculating the real coordinates of the acquired point cloud data based on its calibration data, performing dedistortion processing, and then performing registration based on the data of the base map to obtain the fused point cloud data.

[0017] Furthermore, the data registration includes calculating the real coordinates of the first point cloud data and the second point cloud data obtained based on their respective calibration data, de-distorting them, and then performing registration based on the data of the base map to obtain the fused point cloud data.

[0018] Optionally, the filtering of ground point cloud from the fused point cloud data based on the base map data includes denoising.

[0019] Furthermore, the method further includes: receiving image data collected by a visual sensor, wherein the image data is used for fusion with the point cloud data or for remote video monitoring.

[0020] Furthermore, extracting the minimum bounding box of the at least one subset containing a single target to obtain the perception result includes: calculating the parameters of the minimum bounding box of the at least one subset containing a single target; filtering out false detection results according to the parameters of the minimum bounding box to obtain the perception result.

[0021] Furthermore, the method also includes: sending the perception result to an unmanned vehicle and / or a cloud platform.

[0022] A second aspect of the present disclosure provides a perception system for roadside equipment, comprising:

[0023] A base map making module, used for making topographic base map data of the area where the roadside equipment is located;

[0024] A receiving module, configured to receive point cloud data acquired by at least one laser radar sensor provided on the roadside equipment;

[0025] a fusion module, configured to perform data registration on the base map data and the point cloud data acquired by the at least one lidar sensor to obtain fused point cloud data;

[0026] a segmentation module, configured to filter out ground point clouds from the fused point cloud data based on the base map data, and then segment the data after filtering out the ground point clouds to obtain at least one subset containing a single target;

[0027] The processing module is used to extract a minimum bounding box for the at least one subset containing a single target, and obtain a perception result within the area.

[0028] Furthermore, the system includes an unmanned driving computing platform arranged in a mobile device, and the unmanned driving computing platform includes the fusion module, the segmentation module and the processing module.

[0029] Furthermore, the system includes a mounting rod detachably arranged on the movable device, the at least one laser radar sensor is arranged on the mounting rod, and the mounting rod is retractable or foldable.

[0030] Optionally, the at least one lidar sensor includes a main lidar sensor for acquiring first point cloud data and a blind-complement lidar sensor for acquiring second point cloud data; the fusion module is used to perform data registration on the base map data with the first point cloud data and the second point cloud data respectively to obtain the fused point cloud data.

[0031] Furthermore, the system also includes a communication unit, which is used to send the perception results to the unmanned vehicle and / or the cloud platform.

[0032] Optionally, the communication unit includes a V2V and / or V2N communication unit.

[0033] Optionally, the communication unit can download the base map data from a cloud platform.

[0034] Furthermore, the system also includes a visual sensor, and the image data collected by the visual sensor is used for fusion with the point cloud data or for remote video monitoring.

[0035] Optionally, a fixed pile is provided at the bottom of the movable device.

[0036] Optionally, the movable device is a roadside vehicle.

[0037] According to a third aspect of the present disclosure, an electronic device is provided, including:

[0038] one or more processors;

[0039] A storage unit is used to store one or more programs, which, when executed by the one or more processors, enable the one or more processors to implement the aforementioned perception method of the roadside equipment.

[0040] A fourth aspect of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program can implement the aforementioned perception method of the roadside equipment when executed by a processor.

[0041] A perception method and system for roadside equipment in an embodiment of the present disclosure uses multiple sensors to obtain unmanned driving environment data, and aligns and fuses the data of multiple sensors based on true base map data to obtain real-time perception results, and sends them to the autonomous driving unmanned vehicle side or the cloud. This meets the perception requirements of the open-pit mine autonomous driving system for roads and targets outside the system, assists in the planning of the autonomous driving system, and ensures the safety, stability and operational efficiency of the autonomous driving system. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] FIG1 is a schematic structural diagram of a roadside sensing device according to an embodiment of the present disclosure;

[0043] FIG2 is a schematic flow chart of a roadside sensing method according to an embodiment of the present disclosure;

[0044] FIG3 is a schematic diagram of an application scenario of a roadside perception method according to an embodiment of the present disclosure;

[0045] FIG4 is a structural block diagram of a roadside perception system according to an embodiment of the present disclosure;

[0046] FIG5 is a schematic structural diagram of a roadside perception system according to an embodiment of the present disclosure. Implementation Method

[0047] The following will be combined with the accompanying drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. It should be noted that the embodiments described in the specific embodiments of the present disclosure are merely examples of the embodiments of the present disclosure. Based on the embodiments in the specific embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the embodiments of the present disclosure.

[0048] FIG1 is a schematic diagram of the structure of a roadside perception device 100 provided in an embodiment of the present disclosure. The roadside perception device 100 includes a mobile device 101, an unmanned driving computing platform 102 disposed within the mobile device 101, and at least one lidar sensor, such as a main lidar sensor 103, a blind spot lidar sensor 104, an image sensor 105, a communication unit 106, and a positioning antenna 107. The mobile device 101 can be a roadside vehicle or any other movable device.

[0049] It should be noted that in the embodiments of the present disclosure, there is no restriction on the number of lidar sensors, main lidar sensors, and / or blind-complement lidar sensors. Those skilled in the art can freely choose the number of lidar sensors, main lidar sensors, and / or blind-complement lidar sensors on the premise that they can implement the technical solutions of the present disclosure.

[0050] The unmanned driving computing platform 102 is electrically connected to the main lidar sensor 103, the blind spot-filling lidar sensor 104, the image sensor 105, the communication unit 106, and the positioning antenna 107. The main lidar sensor 103 and the blind spot-filling lidar sensor 104 are respectively configured to obtain first point cloud data and second point cloud data of the unmanned driving environment. The unmanned driving computing platform 102 is configured to fuse the first and second point cloud data and obtain real-time perception results based on the fusion results.

[0051] The image sensor 105 is a visual sensor used to acquire real-time images and / or video data of the environment. The visual sensor is preferably a camera or a video camera. The real-time images and / or video data acquired by the image sensor 105 are used for video monitoring of the environment to assist unmanned driving.

[0052] In a preferred embodiment, the unmanned driving computing platform 102 can also fuse the first point cloud data obtained by the main lidar sensor 103, the second point cloud data obtained by the blind spot lidar sensor 104, and the real-time image and / or video data obtained by the image sensor 105, and obtain real-time perception results based on the fusion results.

[0053] The communication unit 106 is configured to send the real-time perception result to the autonomous driving vehicle and / or the cloud. The communication unit 106 includes a V2V communication module and a V2N communication module.

[0054] The positioning antenna 107 is used to obtain real-time positioning information of the roadside vehicle. For example, GPS or Beidou navigation positioning information, or IMU inertial navigation positioning can be used.

[0055] FIG2 is a flow chart of a roadside sensing method 200 provided in one embodiment of the present disclosure, including:

[0056] Step 201: Create a base map.

[0057] In step 201, radar point cloud data of a target area, such as an intersection, is collected by a vehicle or drone and rasterized to produce a rasterized terrain base map. This rasterized terrain base map serves as a ground truth to assist with data dedistortion and registration in the subsequent point cloud mapping process, as well as ground point cloud filtering in the subsequent target segmentation algorithm, thereby improving the accuracy and efficiency of the acquired perception results.

[0058] The base map data of the topographic base map is saved as a configuration file, and only needs to be produced once when the topography does not change much, or updated at a predetermined time period, such as three months or six months.

[0059] Step 202 , respectively receiving first point cloud data acquired by a first laser radar and second point cloud data acquired by a second laser radar.

[0060] Specifically, FIG3 illustrates an example application scenario of this embodiment, including an intersection formed by four roads: A, B, C, and D. A and B are manned roads outside the autonomous driving system, while C and D are unmanned roads within the system. The roadside sensing device 100 of this embodiment is placed at point RP in FIG3 . Roads A and B are farther from the roadside sensing device, requiring a smaller angle of observation. Roads C and D are closer, requiring a larger angle of observation.

[0061] Referring to Figures 1 to 3 , in this embodiment, the main LiDAR sensor 103 is used to sense roads A and B to obtain the first point cloud data, and the blind spot-filling LiDAR sensor 104 is used to sense roads C and D to obtain the second point cloud data. The detection range of the main LiDAR sensor 103 is greater than the detection range of the blind spot-filling LiDAR sensor 104, the scanning beam of the main LiDAR sensor 103 is greater than the scanning beam of the blind spot-filling LiDAR sensor 104, and the field of view of the main LiDAR sensor 103 is smaller than the field of view of the blind spot-filling LiDAR sensor 104.

[0062] As a specific example, the main LiDAR sensor 103 used to sense roads A and B has a scanning beam of 64 or more lines, for example, 160 lines, an effective detection range of 150 meters or more, for example, 200 meters, and a field of view of 120° or less. The blind spot-filling LiDAR sensor 104 used to scan roads C and D has a scanning beam of less than 32 lines, an effective detection range of less than 50 meters, and a field of view of 360°. As an example, the Tudatong Falcon can be used as the main LiDAR sensor 103, and the Hesai XT32 can be used as the blind spot-filling LiDAR sensor 104.

[0063] In step 203, the real coordinates of the first point cloud data acquired by the main laser radar 103 and the second point cloud data acquired by the blind spot laser radar 104 are calculated based on their respective calibration data, and are aligned based on the data of the base map to obtain fused point cloud data.

[0064] In step 203, the first point cloud data acquired by the main lidar 103 and the second point cloud data acquired by the blind spot filling lidar 104 include the lidar's pitch angle, rotation angle, and depth distance values. The first point cloud data acquired by the main lidar 103 and the second point cloud data acquired by the blind spot filling lidar 104 can be converted into coordinates in a Cartesian coordinate system or a polar coordinate system, and the coordinate values ​​can be aligned with the base map data to obtain accurate fused point cloud data.

[0065] In step 204 , the ground point cloud is filtered out from the fused point cloud data obtained in step 203 , and data segmentation is performed on the data after filtering out the ground point cloud to obtain at least one subset containing a single target.

[0066] In step 204, the fused point cloud data obtained in step 203 is first filtered out of the ground point cloud based on the base map data, and then a series of denoising processes are performed, such as radius filtering, statistical filtering, histogram filtering, etc., to filter out most of the noise, and finally the remaining scattered point sets are divided into subsets containing only individual targets. Commonly used data segmentation algorithms include the region growing method, which achieves clustering by selecting seed points for iterative growth; or the Mean Shift algorithm, which achieves local clustering of the density of the point set by iteratively adjusting the position of the sphere. It should be noted that the embodiments of the present disclosure do not limit the data segmentation algorithm, and those skilled in the art can use other data segmentation methods in the field to achieve the above-mentioned data segmentation.

[0067] In step 205 , for the at least one subset containing a single target, a minimum bounding box is extracted and the minimum bounding box parameters are calculated. Under the premise of ensuring safety and no missed detection, some false detection results are filtered out based on the minimum bounding box parameters to obtain the final perception result.

[0068] Furthermore, the method may also include a communication step (not shown in the figure), sending the above-mentioned perception results to the vehicle side (not shown in the figure) through V2V to assist in unmanned driving path planning and ensure the safety and operational efficiency of the autonomous driving system; or sending the above-mentioned perception results to the cloud platform (not shown in the figure) through V2N to assist the cloud platform in monitoring and scheduling.

[0069] FIG4 is a schematic diagram of the structure of a roadside perception system 400 according to an embodiment of the present disclosure. The roadside perception system 400 includes:

[0070] A base map making module 401 is used to make a base map;

[0071] A receiving module 402 is configured to receive first point cloud data acquired by a first laser radar and second point cloud data acquired by a second laser radar;

[0072] A fusion module 403 is configured to calculate the real coordinates of the first point cloud data acquired by the first laser radar and the second point cloud data acquired by the second laser radar based on their respective calibration data, and to align them based on the data of the base map to obtain fused point cloud data;

[0073] The segmentation module 404 filters out the ground point cloud from the fused point cloud data obtained by the fusion module 403 and performs data segmentation on the data after filtering out the ground point cloud to obtain at least one subset containing a single target;

[0074] Processing module 405 extracts a minimum bounding box for the at least one subset containing a single target, calculates parameters of the minimum bounding box, and filters out some false detections based on the minimum bounding box parameters while ensuring safety and no missed detections, thereby obtaining a final perception result.

[0075] Communication module 406 is used to send the above-mentioned perception results to the vehicle side through V2V to assist in unmanned driving path planning and ensure the safety and operational efficiency of the autonomous driving system; or to send the above-mentioned perception results to the cloud platform through V2N to assist the cloud platform in monitoring and scheduling.

[0076] For example, FIG5 is a schematic diagram of the structure of a roadside sensing device 500 according to an embodiment of the present disclosure. The roadside sensing device 500 includes a roadside vehicle 501, an unmanned driving computing platform 502 mounted on the roadside vehicle 501, and a first supporting bracket 505, a second supporting bracket 506, and a third supporting bracket 507 mounted on a mounting pole 504 in ascending order from the vehicle platform 503.

[0077] The first supporting bracket 505 is equipped with a blind spot laser radar sensor 509, the second supporting bracket 506 is equipped with a main laser radar sensor 508 and an image sensor 510, and the third supporting bracket 507 is equipped with a positioning antenna 512. A communication unit 511 is arranged below the first supporting bracket 505, close to the unmanned computing platform 502.

[0078] Exemplarily, the first supporting bracket 505, the second supporting bracket 506, and the third supporting bracket 507 are arranged in order from low to high. The blind spot compensation lidar sensor 509 is arranged on the first supporting bracket 505, and the main lidar sensor 508 and the image sensor 510 are arranged on the second supporting bracket 506. In this way, the main lidar sensor 508 and the blind spot compensation lidar sensor 509 can be arranged at a relative height. In this embodiment, the blind spot compensation lidar sensor 509 can be a mechanically rotating lidar sensor, so that the 360-degree rotating lidar sensor 509 can be used to assist the main lidar sensor 508 to obtain more comprehensive environmental point cloud data.

[0079] The positioning antenna 512 is disposed on the relatively highest third supporting bracket 507 and is used to transmit the position information of the roadside sensing device 500 to a remote monitoring device, so as to facilitate remote operators to monitor and adjust the position of the roadside sensing device 500.

[0080] Illustratively, the blind spot-filling lidar sensor 509 is disposed on the side of the first supporting bracket 505 facing the second supporting bracket 506. The image sensor 510 is disposed on the side of the second supporting bracket 506 facing the first supporting bracket 505, and the main lidar sensor 508 is disposed on the side of the second supporting bracket 506 facing the third supporting bracket 507. The roadside perception device 500 of this embodiment, by providing the image sensor 510, utilizes image data acquired by the image sensor 510 to monitor the intersection area and process auxiliary point cloud data, further improving perception accuracy and thereby enhancing the safety and reliability of the autonomous driving system.

[0081] For example, as shown in Figure 5, the vehicle platform 503 is disposed on one side of the roadside vehicle 501, and the first load-bearing bracket 505, the second load-bearing bracket 506, and the third load-bearing bracket 507 are all disposed on the other side of the mounting pole 504 facing the roadside vehicle 501. This arrangement allows the ground projection of the center of gravity of the roadside sensing device 500 to be located within the ground projection of the load-bearing portion of the roadside sensing device 500, thereby increasing the overall stability of the roadside device.

[0082] For example, as shown in FIG5 , a fixed stake 513 is provided at the bottom of the vehicle platform 503. When the roadside sensing device 500 is stationary at a predetermined sensing location, the fixed stake 513 is fixed to the ground, thereby fixing the position of the roadside sensing device 500 and enhancing its stability. When the roadside sensing device 500 needs to be moved, the fixed stake 513 is suspended above the ground to facilitate movement.

[0083] 5 , the mounting rod 504 is detachable relative to the vehicle platform 503. Preferably, the mounting rod 504 is retractable or foldable to facilitate its installation and height adjustment.

[0084] The perception method and system for roadside equipment provided in the embodiments of the present disclosure obtain unmanned driving environment data by using multiple sensors, and perform registration and fusion calculations on the data of multiple sensors based on true value base map data to obtain real-time perception results, and send them to the autonomous driving unmanned vehicle side or the cloud. This meets the perception requirements of the open-pit mine autonomous driving system for roads and targets outside the system, assists in the planning of the autonomous driving system, and ensures the safety, stability and operational efficiency of the autonomous driving system.

[0085] An embodiment of the present disclosure also provides an electronic device (not shown in the figure), including a processor and a storage unit, wherein the storage unit is used to store one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors can implement the perception method of the roadside equipment of the embodiment of the present disclosure.

[0086] It should be noted that the division of units in the embodiments of the present disclosure is illustrative and is merely a logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the embodiments of the present disclosure may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0087] The present disclosure also provides a computer-readable storage medium having a program stored thereon. When executed by a processor, the program implements each process of the above-described roadside sensing method embodiment and achieves the same technical effect. To avoid repetition, the program is not described here. The computer-readable storage medium can be any available medium or data storage device accessible by the processor, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MOs), etc.), optical storage (such as CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (such as ROMs, EPROMs, EEPROMs, non-volatile memories (NAND FLASH), solid-state drives (SSDs), etc.).

[0088] It should be noted that the embodiments of the present disclosure do not strictly limit the number and position of each component. On the premise that those skilled in the art can implement the technical solutions of the present disclosure, they can freely select the number and position of each component.

[0089] It is understood that the above embodiments are merely exemplary embodiments for illustrating the principles of the present disclosure, and the present disclosure is not limited thereto. Those skilled in the art may make various modifications and improvements without departing from the spirit and substance of the present disclosure, and such modifications and improvements are also considered to be within the scope of protection of the present disclosure.

Claims

1. A method for sensing roadside equipment, characterized in that: The method comprises: Producing topographic base map data of the area where the roadside equipment is located; Receiving point cloud data acquired by at least one laser radar sensor disposed on the roadside equipment; Performing data registration on the base map data and the point cloud data acquired by the at least one laser radar sensor to obtain fused point cloud data; Filtering the ground point cloud from the fused point cloud data based on the base map data, and then segmenting the data after filtering the ground point cloud to obtain at least one subset containing a single target; A minimum bounding box is extracted for the at least one subset containing a single target to obtain a perception result within the area.

2. The method according to claim 1, characterized in that The at least one laser radar sensor includes a main laser radar sensor for acquiring first point cloud data and a blind laser radar sensor for acquiring second point cloud data; The registering the base map data with the point cloud data acquired by the at least one laser radar sensor includes fusing the first point cloud data with the second point cloud data.

3. The method according to claim 2, characterized in that The main laser radar sensor and the blind spot laser radar sensor have different detection distances or field of view angles.

4. The method according to claim 3, characterized in that The scanning line beam of the main laser radar sensor is larger than the scanning line beam of the blind-complementing laser radar sensor, and the detection distance of the main laser radar is larger than the detection distance of the blind-complementing laser radar sensor.

5. The method according to claim 4, characterized in that The scanning beam of the main laser radar sensor is more than 64 lines, the detection distance is more than 150 meters, and the scanning angle is less than 120°; the scanning beam of the blind spot laser radar sensor is less than 32 beams, the detection distance is less than 50 meters, and the scanning angle is 360°.

6. The method according to claim 1, characterized in that The data registration includes calculating the real coordinates of the acquired point cloud data based on its calibration data, performing de-distortion processing, and then performing registration based on the data of the base map to obtain the fused point cloud data.

7. The method according to claim 2, characterized in that: The data registration includes calculating the real coordinates of the first point cloud data and the second point cloud data obtained based on their respective calibration data, performing de-distortion processing, and then performing registration based on the data of the base map to obtain the fused point cloud data.

8. The method according to claim 1 or 2, characterized in that: The filtering of ground point cloud from the fused point cloud data based on the base map data includes denoising.

9. The method according to claim 1 or 2, characterized in that: The method further includes: receiving image data collected by a visual sensor, wherein the image data is used for fusion with the point cloud data or for remote video monitoring.

10. The method according to claim 1 or 2, characterized in that: The extracting of the minimum bounding box of the at least one subset containing the single target to obtain the perception result includes: calculating the parameters of the minimum bounding box of the at least one subset containing the single target; filtering out false detection results according to the parameters of the minimum bounding box to obtain the perception result.

11. The method according to claim 1 or 2, characterized in that: The method also includes: sending the perception result to an unmanned vehicle and / or a cloud platform.

12. A perception system for roadside equipment, characterized in that: The system comprises: A base map making module, used to make topographic base map data of the area where the roadside equipment is located; A receiving module, used to receive point cloud data acquired by at least one laser radar sensor arranged on the roadside equipment; A fusion module, used for performing data registration on the base map data and the point cloud data acquired by the at least one laser radar sensor to obtain fused point cloud data; A segmentation module, used for filtering the ground point cloud from the fused point cloud data based on the base map data, and then performing data segmentation on the data after filtering the ground point cloud to obtain at least one subset containing a single target; The processing module is used to extract a minimum bounding box for the at least one subset containing a single target to obtain a perception result within the area.

13. The system according to claim 12, characterized in that The system includes an unmanned driving computing platform arranged in a mobile device, and the unmanned driving computing platform includes the fusion module, the segmentation module and the processing module.

14. The system according to claim 13, characterized in that The system includes a mounting pole detachably arranged on the movable device, the at least one laser radar sensor is arranged on the mounting pole, and the mounting pole is retractable or foldable.

15. The system according to claim 14, characterized in that The at least one laser radar sensor includes a main laser radar sensor for acquiring first point cloud data and a blind laser radar sensor for acquiring second point cloud data; The fusion module is used to perform data registration on the base map data with the first point cloud data and the second point cloud data respectively to obtain the fused point cloud data.

16. The system according to claim 12 or 15, characterized in that The system also includes a communication unit, which is used to send the perception result to the unmanned vehicle and / or the cloud platform.

17. The system according to claim 16, characterized in that The communication unit includes a V2V and / or V2N communication unit.

18. The system according to claim 16, characterized in that The communication unit is capable of downloading the base map data from a cloud platform.

19. The system according to claim 12 or 15, characterized in that The system also includes a visual sensor, and image data collected by the visual sensor is used for fusion with the point cloud data or for remote video monitoring.

20. The system according to claim 14 or 15, characterized in that A fixing pile is arranged at the bottom of the movable device.

21. The system according to claim 14 or 15, characterized in that The movable device is a roadside vehicle.

22. An electronic device, characterized in that: include: one or more processors; A storage unit, used to store one or more programs, which, when executed by the one or more processors, enable the one or more processors to implement the perception method of the roadside equipment according to any one of claims 1 to 11.

23. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the perception method of the roadside equipment according to any one of claims 1 to 11 can be implemented.

Images