Parking lot obstacle hanging detection method and system, storage medium and vehicle

Abstract

The application discloses a detection method for a suspended obstacle in a parking lot. The method comprises the following steps: acquiring point cloud data of the parking lot environment by a laser radar loaded on a vehicle body; performing corresponding processing, such as screening, plane fitting, clustering operation and solving a minimum circumscribed cube, on the point cloud data; and obtaining a characteristic parameter of the minimum circumscribed cube, so as to realize detection of the suspended obstacle in the parking lot. The application also provides a corresponding system, a storage medium and a vehicle. The application can realize detection of the suspended obstacle in the parking lot, and has the advantages of easy implementation, low cost, high accuracy, good stability and strong adaptability.

Description

Technical Field

[0001] This invention relates to the technical field of obstacle detection, and in particular to a method, system, storage medium, and vehicle for detecting suspended obstacles in a parking lot. Background Technology

[0002] Obstacle detection in parking lots is a crucial part of environmental perception in automated parking systems. Currently, the main technologies for detecting obstacles in parking lots include visual detection and ultrasonic detection.

[0003] Vision-based obstacle detection is primarily accomplished by cameras surrounding the vehicle. These cameras capture images of the surrounding environment, which are then processed using relevant algorithms to detect obstacles. Currently, there are two main types of image processing algorithms: one is a traditional image processing-based visual detection method; the other is a deep learning-based visual detection method, which requires significant computational power, consumes substantial computing resources, and is not easily interpretable. Overall, visual detection methods are relatively inexpensive. However, both traditional image processing-based and deep learning-based methods are sensitive to lighting conditions. Recognition performance is significantly affected by complex lighting environments, and accuracy drops drastically in low-light conditions.

[0004] Ultrasonic obstacle detection utilizes ultrasound as a signal source. An ultrasonic sensor emits ultrasonic waves; when these waves encounter an obstacle, some are reflected back. A receiver picks up the reflected waves, and the sensor calculates the travel time of the reflected waves in the air using a timer. Based on this travel distance, the distance between the obstacle and the sensor is calculated, thus enabling obstacle detection. However, the speed of sound is easily affected by weather conditions; temperature changes affect the speed of sound, increasing measurement errors. Ultrasonic echoes are weak at long distances, making them unsuitable for detecting and tracking obstacles at greater distances. Furthermore, ultrasonic radar, typically mounted on vehicle bumpers, has a small measurement angle and short range, while visual sensors, although with longer detection ranges, are susceptible to light interference and cannot effectively identify distant, high-positioned suspended obstacles.

[0005] Currently, no solution has been found that is low-cost, stable, and adaptable to various environments for detecting suspended obstacles in parking lots. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method, system, storage medium and vehicle for detecting suspended obstacles in parking lots, which can detect suspended obstacles in parking lots and has the advantages of being easy to implement, low cost, high accuracy, good stability and strong adaptability.

[0007] To address the aforementioned technical problems, as one aspect of the present invention, a method for detecting suspended obstacles in a parking lot is provided, which includes at least the following steps:

[0008] Point cloud data of the surrounding environment is acquired by an onboard LiDAR installed on the roof of the vehicle;

[0009] Among the acquired point cloud data, retain the point cloud data of the predetermined region of interest;

[0010] Perform plane fitting on the point cloud data of the region of interest to obtain multiple planes, and remove the point cloud data corresponding to the plane with the largest area.

[0011] Cluster the remaining point cloud data after removing the plane with the largest area to obtain at least one cluster category, and solve for the minimum bounding cube corresponding to the point cloud data of each cluster category.

[0012] Obtain the representation parameters of the minimum bounding cube corresponding to the point cloud data of each cluster category, and output them as the detection results of the suspended obstacle.

[0013] Prior to performing planar fitting on the point cloud data of the region of interest, the process further includes:

[0014] The point cloud data of the region of interest is filtered to remove noise.

[0015] The step of retaining point cloud data of a predetermined region of interest from the acquired point cloud data further includes:

[0016] Convert the coordinates of each point cloud data into coordinates in its vehicle body coordinate system to obtain the distance between each point cloud data and the vehicle.

[0017] Based on the distance between each point cloud data and the vehicle, it is compared with a pre-set region of interest, and all point cloud data outside the region of interest are excluded, while point cloud data within the region of interest are retained.

[0018] The region of interest is defined as the area consisting of a distance of 0.5-2.5 meters above the ground, ±2 meters to both sides of the vehicle's center, and 20 meters in front of the vehicle.

[0019] The characterization parameters include at least the center point coordinates, length, width, height, and deflection angle.

[0020] Accordingly, another aspect of the present invention provides a detection system for suspended obstacles in a parking lot, comprising at least:

[0021] The point cloud data acquisition unit is used to acquire point cloud data of the surrounding environment through an onboard LiDAR installed on the roof of the vehicle.

[0022] The filtering and processing unit is used to retain point cloud data of a predetermined region of interest from the acquired point cloud data;

[0023] The plane fitting processing unit is used to perform plane fitting on the point cloud data of the region of interest, obtain multiple planes, and remove the point cloud data corresponding to the plane with the largest area.

[0024] The clustering processing unit is used to perform clustering operations on the remaining point cloud data after removing the plane with the largest area, to obtain at least one cluster category, and to solve for the minimum bounding cube corresponding to the point cloud data of each cluster category.

[0025] The detection result output unit is used to obtain the representation parameters of the minimum bounding cube corresponding to the point cloud data of each cluster category, and output them as the detection result of the suspended obstacle.

[0026] This further includes:

[0027] The noise reduction unit is used to filter the point cloud data of the region of interest and remove noise data.

[0028] The filtering processing unit further includes:

[0029] The coordinate transformation unit is used to convert the coordinates of each point cloud data into coordinate values ​​in its vehicle body coordinate system, thereby obtaining the distance value between each point cloud data and the vehicle.

[0030] The comparison processing unit is used to compare each point cloud data with a pre-set region of interest based on the distance value between each point cloud data and the vehicle, exclude all point cloud data outside the region of interest, and retain the point cloud data within the region of interest.

[0031] The region of interest is defined as the area consisting of a distance of 0.5-2.5 meters above the ground, ±2 meters to both sides of the vehicle's center, and 20 meters in front of the vehicle.

[0032] The characterization parameters include at least the center point coordinates, length, width, height, and deflection angle.

[0033] Accordingly, another aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned method.

[0034] Accordingly, another aspect of the present invention provides a vehicle with an onboard lidar mounted on its roof, which integrates a parking lot obstacle detection system.

[0035] Implementing the embodiments of the present invention has the following beneficial effects:

[0036] This invention provides a method, system, storage medium, and vehicle for detecting suspended obstacles in parking lots. By using point cloud data of suspended obstacles in parking lots obtained by LiDAR, operations such as ROI extraction, plane fitting, plane deletion, clustering, and minimization of cubes are performed to effectively detect the location of suspended obstacles on walls. This method does not require the collection and labeling of large amounts of data and has low computational requirements, thus providing a solid foundation for ensuring the smooth operation of automatic parking systems.

[0037] Meanwhile, the method of the present invention has excellent stability and is not affected by environmental factors such as light or temperature, unlike visual sensors or ultrasonic sensors. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, obtaining other drawings based on these drawings without creative effort still falls within the scope of the present invention.

[0039] Figure 1 This is a schematic diagram of the main flow of an embodiment of a method for detecting suspended obstacles in a parking lot provided by the present invention;

[0040] Figure 2 This is a schematic diagram of a structure of an embodiment of a parking lot hanging obstacle detection system provided by the present invention;

[0041] Figure 3 for Figure 2 A schematic diagram of the structure of the screening and processing unit. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0043] like Figure 1 The diagram shows a main flowchart of an embodiment of a method for detecting suspended obstacles in a parking lot provided by the present invention; in this embodiment, the method includes at least the following steps:

[0044] Step S10 involves acquiring point cloud data of the surrounding environment using an onboard LiDAR mounted on the vehicle roof. Since the LiDAR is mounted on the roof, it can scan all suspended obstacles above and below the vehicle's height into the point cloud data. It's understood that LiDAR detects the distance to obstacles by transmitting and receiving laser beams; its detection process is unaffected by environmental factors such as light and temperature, exhibiting excellent stability.

[0045] Step S11: In the acquired point cloud data, retain the point cloud data of the predetermined region of interest;

[0046] In one specific embodiment, step S11 further includes:

[0047] Step S110: Convert the coordinates of each point cloud data into coordinates in the vehicle body coordinate system to obtain the distance between each point cloud data and the vehicle. It can be understood that since the LiDAR is fixed on the roof of the vehicle, it is necessary to calibrate to convert the x, y, z of each point in the point cloud data into the actual distance of the target point in the vehicle body coordinate system.

[0048] Step S111: Based on the distance value between each point cloud data and the vehicle, compare it with a pre-set region of interest (ROI), exclude all point cloud data outside the region of interest, and retain the point cloud data within the region of interest;

[0049] Understandably, in this invention, only suspended obstacles that may affect the vehicle need to be considered; obstacles that are too far away or too high from the vehicle do not need to be considered. Suspended obstacles refer to protruding obstacles fixed to a wall at a certain height, such as fire hydrants and electrical boxes. Since suspended obstacles have a certain height, a range of ground clearance can be set. Areas too far from the sides of the vehicle do not affect vehicle operation and therefore do not need to be considered; thus, predetermined positive and negative ROI areas can be set from the center of the vehicle body to both sides. In parking scenarios, due to the slow vehicle speed, obstacles too far in front of the vehicle do not need to be considered; a predetermined area in front can be set as the Region of Interest (ROI). By setting the ROI areas as described above, a large number of irrelevant areas can be eliminated, thereby significantly reducing the difficulty of recognition and the computational load.

[0050] Specifically, in one embodiment, the region of interest can be defined as the area consisting of a distance of 0.5-2.5 meters above the ground, ±2 meters to both sides of the vehicle center, and 20 meters in front of the vehicle.

[0051] Step S12: Perform plane fitting on the point cloud data of the region of interest to obtain multiple planes, and remove the point cloud data corresponding to the plane with the largest area. It is understood that in the embodiments of the present invention, professional point cloud data processing software (such as CloudCompare) can be used to perform plane fitting on the point cloud data.

[0052] Since suspended obstacles are fixed to the wall, and the wall is a plane, its area is obviously larger than the area of ​​the obstacle. When performing planar fitting on the ROI region, at least the plane containing the wall will be fitted. When the obstacle is a cube like a fire hydrant, its surface will also be fitted by the planar fitting operation. This will result in multiple fitted planes, with the largest plane being the wall. Removing this largest plane leaves only the point cloud information of the obstacle mounted on the wall.

[0053] Step S13: Perform clustering operation on the remaining point cloud data after removing the plane with the largest area to obtain at least one cluster category, and solve for the minimum bounding cube corresponding to the point cloud data of each cluster category.

[0054] Understandably, after removing the wall, the obstacle point cloud information that was originally fixed at different positions on the wall is separated. At this point, clustering is performed on the point cloud, and then each cluster is enclosed by the smallest bounding cube, which enables the detection of the obstacle's position.

[0055] It is understood that in this invention, the clustering operation can employ clustering algorithms such as K-means and density-based spatial clustering (DBSCAN). There are also mature implementation methods and algorithms for calculating the minimum bounding cube from point cloud data; for example, the Moving Cubes (MC) function module in the Point Cloud Library (PCL) software can be used to obtain the minimum bounding cube data.

[0056] Step S14: Obtain the representation parameters of the minimum bounding cube corresponding to the point cloud data of each cluster category, and output them as the detection result of the suspended obstacle. The representation parameters include at least the center point coordinates, length, width, height, and deflection angle of the obstacle.

[0057] In step S12, before performing planar fitting on the point cloud data of the region of interest, the following further steps are included:

[0058] The point cloud data of the region of interest is filtered to remove noise. This step removes outliers or noisy data, improving the plane fitting and clustering results in subsequent steps and increasing computational efficiency.

[0059] like Figure 2 The diagram shown illustrates a structural schematic of an embodiment of a parking lot obstacle detection system provided by the present invention. (In conjunction with...) Figure 3 As shown, the parking lot hanging obstacle detection system includes at least:

[0060] The point cloud data acquisition unit 10 is used to acquire point cloud data of the surrounding environment through an onboard LiDAR installed on the roof of the vehicle.

[0061] The filtering processing unit 11 is used to retain point cloud data of a predetermined region of interest in the acquired point cloud data;

[0062] The plane fitting processing unit 13 is used to perform plane fitting on the point cloud data of the region of interest, obtain multiple planes, and remove the point cloud data corresponding to the plane with the largest area.

[0063] Clustering processing unit 14 is used to perform clustering operations on the remaining point cloud data after removing the plane with the largest area, to obtain at least one cluster category, and to solve the minimum bounding cube corresponding to the point cloud data of each cluster category.

[0064] The detection result output unit 15 is used to obtain the representation parameters of the minimum bounding cube corresponding to the point cloud data of each cluster category, and output them as the detection result of the suspended obstacle. The representation parameters include at least: center point coordinates, length, width, height, and deflection angle.

[0065] This further includes:

[0066] The noise reduction unit 12 is used to filter the point cloud data of the region of interest and remove noise data.

[0067] In a specific example, the filtering processing unit 11 further includes:

[0068] The coordinate transformation unit 110 is used to convert the coordinates of each point cloud data into coordinate values ​​in its vehicle body coordinate system, so as to obtain the distance value between each point cloud data and the vehicle.

[0069] The comparison processing unit 111 is used to compare each point cloud data with a pre-set region of interest based on the distance value between each point cloud data and the vehicle, exclude all point cloud data outside the region of interest, and retain the point cloud data within the region of interest.

[0070] The region of interest is defined as the area consisting of a distance of 0.5-2.5 meters above the ground, ±2 meters to both sides of the vehicle's center, and 20 meters in front of the vehicle.

[0071] For more details, please refer to and combine with the above. Figure 1The description of that will not be repeated here.

[0072] Accordingly, in another aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned... Figure 1 The steps of the described method. For more details, please refer to and combine with the foregoing descriptions. Figure 1 The description of that will not be repeated here.

[0073] Accordingly, another aspect of the present invention also provides a vehicle having an onboard lidar mounted on its roof, which integrates, for example... Figures 2 to 3 The described parking lot obstacle detection system. For more details, please refer to and combine with the aforementioned... Figures 1 to 3 The description of that will not be repeated here.

[0074] Implementing the embodiments of the present invention has the following beneficial effects:

[0075] This invention provides a method, system, storage medium, and vehicle for detecting suspended obstacles in parking lots. By using point cloud data of suspended obstacles in parking lots obtained by LiDAR, operations such as ROI extraction, plane fitting, plane deletion, clustering, and minimization of cubes are performed to effectively detect the location of suspended obstacles on walls. This method does not require the collection and labeling of large amounts of data and has low computational requirements, thus providing a solid foundation for ensuring the smooth operation of automatic parking systems.

[0076] Meanwhile, the method of the present invention has excellent stability and is not affected by environmental factors such as light or temperature, unlike visual sensors or ultrasonic sensors.

[0077] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0078] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0079] The above description is merely a preferred embodiment of the present invention and should not be construed as limiting the scope of the invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for detecting suspended obstacles in a parking lot, characterized in that, It should include at least the following steps: Step S10: Obtain point cloud data of the surrounding environment by using the vehicle-mounted LiDAR installed on the roof of the vehicle, and scan all suspended obstacles above and below the height of the vehicle body into the point cloud data. Step S11: In the acquired point cloud data, retain the point cloud data of the predetermined region of interest, specifically including: Convert the coordinates of each point cloud data into coordinates in its vehicle body coordinate system to obtain the distance between each point cloud data and the vehicle. Based on the distance between each point cloud data and the vehicle, it is compared with a pre-set region of interest, and all point cloud data outside the region of interest are excluded, while point cloud data within the region of interest are retained; The region of interest is defined as the area consisting of 0.5-2.5 meters above the ground, ±2 meters to both sides of the vehicle's center, and 20 meters in front of the vehicle. Step S12: Filter the point cloud data of the region of interest to remove noise data; Step S13: Perform plane fitting on the point cloud data of the region of interest to obtain multiple planes, and remove the point cloud data corresponding to the plane with the largest area corresponding to the wall. Step S14: Perform clustering on the remaining point cloud data after removing the plane with the largest area to obtain at least one cluster category, and solve for the minimum bounding cube corresponding to the point cloud data in each cluster category; the clustering operation adopts the K-means clustering algorithm or the density-based spatial clustering algorithm. Step S15: Obtain the representation parameters of the minimum bounding cube corresponding to the point cloud data of each cluster category, and output them as the detection result of the suspended obstacle. The representation parameters include at least the center point coordinates, length, width, height and deflection angle.

2. A detection system for suspended obstacles in a parking lot, characterized in that, At least including: The point cloud data acquisition unit is used to acquire point cloud data of the surrounding environment through the vehicle-mounted LiDAR installed on the roof of the vehicle, and to scan all suspended obstacles above and below the vehicle body height into the point cloud data; The filtering and processing unit is used to retain point cloud data of a predetermined region of interest from the acquired point cloud data; the region of interest is the area consisting of a distance of 0.5-2.5 meters above the ground, ±2 meters to both sides of the vehicle center, and 20 meters in front of the vehicle. The noise reduction unit is used to filter the point cloud data of the region of interest and remove noise data; The plane fitting processing unit is used to perform plane fitting on the point cloud data of the region of interest, obtain multiple planes, and remove the point cloud data corresponding to the plane with the largest area corresponding to the wall. The clustering processing unit is used to perform clustering operations on the remaining point cloud data after removing the plane with the largest area, to obtain at least one cluster category, and to solve for the minimum bounding cube corresponding to the point cloud data of each cluster category; the clustering operation adopts the K-means clustering algorithm or the density-based spatial clustering algorithm. The detection result output unit is used to obtain the representation parameters of the minimum bounding cube corresponding to the point cloud data of each cluster category, and output them as the detection result of the suspended obstacle. The representation parameters include at least: center point coordinates, length, width, height and deflection angle. The filtering processing unit further includes: The coordinate transformation unit is used to convert the coordinates of each point cloud data into coordinate values ​​in its vehicle body coordinate system, thereby obtaining the distance value between each point cloud data and the vehicle. The comparison processing unit is used to compare each point cloud data with a pre-set region of interest based on the distance value between each point cloud data and the vehicle, exclude all point cloud data outside the region of interest, and retain the point cloud data within the region of interest.

3. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method as described in claim 1.

4. A vehicle, wherein an on-board lidar is installed on its roof, characterized in that, It further integrates the parking lot hanging obstacle detection system as described in claim 2.

Images