Automatic parking method and device and computer program product

Abstract

The invention discloses an automatic parking method and device and a computer program product, and the method comprises the steps: screening ground point cloud from collected three-dimensional point cloud data of the surrounding environment of a vehicle; performing clustering processing on the ground point cloud to generate a plurality of point clusters; extracting a space boundary parameter of each point cluster, and calculating the length and width of a rectangular region corresponding to each point cluster; based on the size of the vehicle, the rectangular areas with the length and the width larger than the corresponding size of the vehicle are screened out to serve as candidate berthable areas; and parking path planning and automatic parking operation are executed based on a target available area selected by a user from the candidate available areas. Aiming at the inherent defect that a traditional automatic parking system depends on structured parking space characteristics, a three-dimensional terrain flatness quantitative evaluation mechanism is innovatively introduced, and the technical problems of parking area detection and vehicle mechanical protection in an unstructured scene are effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent connected vehicles, and in particular to an automatic parking method, device, and computer program product. Background Art

[0002] With the advancement of autonomous driving technology, automated parking assistance systems have evolved from basic parking space detection to fully automated parking path planning. Existing solutions typically employ a multi-sensor fusion architecture, combining visual sensors, ultrasonic radar, and millimeter-wave radar to construct a three-dimensional topological model of the parking environment. Parking space recognition primarily involves pattern matching of parking space lines based on standardized structural features such as perpendicular and parallel parking spaces, or constructing a spatial parking model through obstacle spacing analysis.

[0003] However, the inventors' research has revealed that existing automated parking assistance functions fail to consider the user's need for surface flatness. Most automated parking assistance functions determine whether an area is available for parking based on basic criteria, such as whether it is a marked regular parking space or a common space. These functions ignore the flatness of the road surface within these areas, as well as the user's need for a flat parking area in wide areas, such as when camping outdoors. After a vehicle is automatically parked, the uneven surface often causes the vehicle's body to remain unevenly loaded for extended periods, leading to risks of frame deformation, suspension system performance degradation, and other degradation in vehicle quality. Summary of the Invention

[0004] The technical problem to be solved by the embodiments of the present invention is to provide an automatic parking method, device and computer program product to reduce the impact of uneven road surface on the vehicle frame, suspension and wheels during the automatic parking process.

[0005] To solve the above technical problems, the present invention provides an automatic parking method, comprising the following steps:

[0006] Filtering ground point clouds from the collected three-dimensional point cloud data of the vehicle's surrounding environment;

[0007] performing clustering processing on the ground point cloud to generate a plurality of point clusters;

[0008] Extract the spatial boundary parameters of each point cluster and calculate the length and width of the rectangular area corresponding to each point cluster;

[0009] Based on the vehicle size, the rectangular area having a length and a width both larger than the corresponding size of the vehicle is selected as a candidate parking area;

[0010] Based on the target parking area selected by the user from the candidate parking areas, parking path planning and automatic parking operations are performed.

[0011] Preferably, clustering the ground point cloud to generate a plurality of point clusters specifically includes:

[0012] The filtered ground point clouds are spatially traversed, and adjacent point clouds that simultaneously meet the following conditions are clustered to obtain several point clusters:

[0013] The Euclidean distance in the horizontal projection plane is ≤ the preset distance threshold;

[0014] The height difference in the z-axis direction is less than or equal to the preset height difference threshold;

[0015] The vehicle's center of mass is taken as the origin, and the z-axis is perpendicular to the horizontal plane.

[0016] Preferably, extracting the spatial boundary parameters of each point cluster and calculating the length and width of the rectangular area corresponding to each point cluster specifically includes:

[0017] Traverse the x and y coordinate values ​​of all points in each point cluster and determine the maximum x-axis coordinate x max , the minimum x-axis coordinate x min , the maximum value of the y-axis coordinate y max , minimum y-axis coordinate y min ;

[0018] Calculate the extreme value spacing on the horizontal projection plane to obtain the rectangular area size of each point cluster, where the length L = x max -x min , width W = y max -y min .

[0019] Preferably, the step of selecting the rectangular area having a length and a width both larger than the corresponding dimensions of the vehicle as a candidate parking area based on the vehicle dimensions specifically includes:

[0020] The length L and width W of the rectangular area corresponding to each point cluster are respectively compared with the vehicle's own length L veh , width W veh Compare and select the ones that satisfy L>L veh And W>W veh The rectangular area is taken as the candidate parking area.

[0021] Preferably, the performing of parking path planning and automatic parking operation based on the target parking area selected by the user from the candidate parking areas specifically includes:

[0022] Outputting the candidate parking area to an onboard human-computer interaction interface;

[0023] Based on a target parking area selected by a user from the candidate parking areas, outputting the coordinates of four corner points of the target parking area, and performing route planning using an automatic parking planning algorithm;

[0024] The vehicle is controlled to park in the target parking area according to the planned route.

[0025] Preferably, the step of filtering out the ground point cloud from the collected three-dimensional point cloud data of the vehicle surrounding environment specifically includes:

[0026] Performing point cloud hierarchical filtering based on a height empirical threshold H on the collected three-dimensional point cloud data of the vehicle's surrounding environment;

[0027] Filtering out the point cloud whose z-axis coordinate satisfies z∈[-H-δ,-H+δ] as the ground point cloud;

[0028] The three-dimensional point cloud data adopts the vehicle center coordinate system, with the positive x-axis pointing in the direction of vehicle travel and the origin located at the vehicle center of mass; the positive y-axis extends perpendicular to the x-axis to the right, and the positive z-axis extends perpendicular to the horizontal plane and upward; δ is the height tolerance parameter.

[0029] Preferably, the method further comprises: performing secondary verification on the ground point cloud in combination with the reflection intensity value in the three-dimensional point cloud data to remove point cloud data of non-ground materials.

[0030] The present invention also provides an automatic parking device, comprising:

[0031] A first screening module is used to screen out ground point clouds from the collected three-dimensional point cloud data of the vehicle's surrounding environment;

[0032] A clustering module, used to perform clustering processing on the ground point cloud to generate a plurality of point clusters;

[0033] The calculation module is used to extract the spatial boundary parameters of each point cluster and calculate the length and width of the rectangular area corresponding to each point cluster;

[0034] A second screening module is configured to screen out, based on the vehicle size, the rectangular areas whose length and width are both larger than the corresponding dimensions of the vehicle as candidate parking areas;

[0035] The control module is configured to execute parking path planning and automatic parking operations based on a target parking area selected by a user from the candidate parking areas.

[0036] The present invention also provides an automatic parking device, comprising:

[0037] one or more processors;

[0038] Memory;

[0039] One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, and the one or more applications are configured to perform the automatic parking method.

[0040] The present invention also provides a computer program product, comprising computer instructions, wherein the computer instructions instruct a computer device to execute operations corresponding to the method.

[0041] The present invention has the following beneficial effects: Addressing the inherent drawback of traditional automated parking systems, which rely on structured parking space features, the present invention innovatively introduces a quantitative assessment mechanism for three-dimensional terrain flatness, effectively addressing the technical challenges of parking area detection and vehicle mechanical protection in unstructured scenarios. By integrating highly layered filtering of LiDAR point clouds with a dual-constraint clustering algorithm, the system accurately identifies parking areas where surface roughness meets the suspension system's safety threshold, overcoming the limitations of traditional solutions that rely solely on two-dimensional plane detection based on parking space lines or obstacle spacing. In non-standard scenarios such as outdoor camping and temporary parking lots, the present invention enables flat area detection with adaptive matching of vehicle length and width, extending parking operations to complex terrain such as grass and sand, and improving its applicability in unstructured scenarios compared to existing technologies. Furthermore, the selection of parking areas based on three-dimensional terrain features significantly reduces the stress concentration caused by long-term parking on sloped or uneven surfaces, reduces the risk of frame torsional deformation, and extends the service life of the suspension system, providing technical support for reliable parking of smart vehicles in specialized scenarios such as wilderness exploration and emergency rescue. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0043] Figure 1 1 is a flow chart of an automatic parking method according to an embodiment of the present invention.

[0044] Figure 2 Schematic diagram of the structure of an automatic parking device according to a second embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following descriptions of the embodiments refer to the accompanying drawings to illustrate specific embodiments in which the present invention may be implemented.

[0046] Please refer to Figure 1 As shown, the first embodiment of the present invention provides an automatic parking method, comprising the following steps:

[0047] Filtering ground point clouds from the collected three-dimensional point cloud data of the vehicle's surrounding environment;

[0048] performing clustering processing on the ground point cloud to generate a plurality of point clusters;

[0049] Extract the spatial boundary parameters of each point cluster and calculate the length and width of the rectangular area corresponding to each point cluster;

[0050] Based on the vehicle size, the rectangular area having a length and a width both larger than the corresponding size of the vehicle is selected as a candidate parking area;

[0051] Based on the target parking area selected by the user from the candidate parking areas, parking path planning and automatic parking operations are performed.

[0052] Through the above steps, it can be seen that the embodiment of the present invention realizes intelligent parking decision based on terrain flatness assessment through ground point cloud screening and clustering processing of three-dimensional point cloud data, combined with rectangular area geometric parameter calculation and dynamic matching of vehicle size, effectively solving the risk of long-term parking damage caused by traditional automatic parking systems ignoring three-dimensional terrain features.

[0053] Specifically, in this embodiment of the present invention, when the user triggers the flat area parking function through the vehicle's human-computer interaction interface, the intelligent driving system activates the multimodal environmental perception module to collect environmental perception data. Specifically, the laser radar collects raw point cloud data with a horizontal field of view angle of ≥120° and a vertical field of view angle of ≥30° around the vehicle at a preset scanning frequency (e.g., 10Hz). This raw point cloud data uses the vehicle center coordinate system, where:

[0054] The positive direction of the x-axis points to the direction of vehicle travel, and the origin is at the center of mass of the vehicle;

[0055] The positive direction of the y-axis is perpendicular to the x-axis and extends to the right, forming a horizontal coordinate system;

[0056] The positive z-axis extends upward perpendicular to the horizontal plane, representing the vertical height dimension;

[0057] Each 3D point cloud data contains spatial coordinates (x, y, z), reflection intensity value and timestamp information.

[0058] After obtaining the three-dimensional point cloud data, the embodiment of the present invention executes a point cloud hierarchical filtering algorithm based on a height empirical threshold H, and filters all point clouds whose height coordinates (i.e., z-axis coordinates) are lower than the height empirical threshold H as ground point clouds.

[0059] The height empirical threshold H can be calculated based on the height H0 of the vehicle center point from the ground (i.e., the vehicle chassis height) and the preset safety margin ΔH:

[0060] H=H0-ΔH

[0061] Where ΔH is the positive compensation value to ensure that the height experience threshold H is slightly lower than the actual vehicle chassis height H0.

[0062] When executing the point cloud hierarchical filtering algorithm based on the height empirical threshold H, the original point cloud is filtered by the z-axis coordinate, retaining the point cloud data that satisfies z∈[-H-δ,-H+δ], where δ is the height tolerance parameter. This filtering operation initially extracts the ground point cloud by eliminating obstacles above the ground (such as curbs and vehicles) and depressions below the ground.

[0063] For example, if the height threshold H = 0.26m and the height tolerance parameter δ = 0.03m, the effective z-axis coordinate range is [-0.26-0.03, -0.26+0.03] = [-0.29, -0.23]. This range is used to filter the ground point cloud beneath the vehicle, preventing non-ground structures such as the bottom edge of the bumper from being misidentified as parking areas.

[0064] Furthermore, embodiments of the present invention provide a reflection intensity-assisted verification mechanism, which combines the reflection intensity values ​​in the 3D point cloud data to perform secondary verification on the initially extracted ground point cloud. For example, the typical reflection intensity threshold for ground materials (such as asphalt and cement) is 20-50, which can effectively distinguish between metal obstacles (reflection intensity > 80) and vegetation (reflection intensity < 15).

[0065] The present invention implements three-dimensional feature clustering of ground point clouds by applying dual constraints: geometric proximity and high-degree continuity between point clouds. Specifically, the selected ground point clouds are spatially traversed, and any two adjacent point clouds that meet both of the following conditions are merged into the same cluster, thereby obtaining multiple point clusters:

[0066] (1) The Euclidean distance between two points on the horizontal projection plane (xy plane) is less than or equal to the preset distance threshold s (the distance threshold s is used to control the spatial extension continuity of the clustering area);

[0067] (2) The height difference between the two points in the vertical direction (z-axis) is less than or equal to a preset height difference threshold h (the height difference threshold h is used to represent the allowable ground undulation tolerance).

[0068] By combining constraints on horizontal spacing and vertical height differences, the system effectively distinguishes boundaries where terrain abrupt changes occur, identifying spatially continuous point clouds with gently varying heights as belonging to the same flat area. This allows the construction of a 3D model of parking areas that meets the safety requirements of vehicle suspension systems. Compared to traditional single-distance clustering algorithms, this embodiment of the present invention can accurately exclude pseudo-flat areas caused by sudden 3D terrain changes, such as steps and ramps, thereby improving the geometric accuracy of parking area detection.

[0069] Next, the three-dimensional space boundary parameters of each clustered point cluster are extracted: the x and y coordinate values ​​of all points in the point cluster are traversed, and the maximum x-axis coordinate x is determined. max , the minimum x-axis coordinate x min , the maximum value of the y-axis coordinate y max , minimum y-axis coordinate y min ; Then, the geometric enclosing rectangle size of the point cluster is obtained by calculating the extreme value spacing on the horizontal projection plane: length L = x max -x min , width W = y max -y min The rectangular area represents the maximum coverage of the point cluster on the horizontal plane, and its length and width parameters reflect the effective space size of the potential parking space. Then the calculated length L and width W are respectively compared with the vehicle's own length L veh , width W veh Compare and select the ones that satisfy L>L veh And W>W veh The rectangular area is taken as the candidate parking area.

[0070] The embodiment of the present invention simplifies complex spatial judgment into two-dimensional geometric dimension verification through boundary modeling of extreme point mapping while retaining the three-dimensional point cloud terrain features, thereby ensuring parking safety margins and significantly reducing computational complexity.

[0071] The intelligent driving system outputs all candidate parking areas to the vehicle human-computer interaction interface and displays them to the user for selection. After the user selects one of the candidate parking areas as the target parking area, the intelligent driving system calculates the four corner points (x max ,y max )、(x max ,y min )、(x min ,y max )、(x min ,y min ) and Hybrd A*, commonly used automatic parking planning algorithms, perform route planning. Finally, a control algorithm is used to control the chassis, powertrain, and other related components to steer the vehicle into the user-specified target parking area.

[0072] For example Figure 2 As shown, corresponding to the automatic parking method described in the first embodiment of the present invention, the second embodiment of the present invention further provides an automatic parking device, including:

[0073] A first screening module is used to screen out ground point clouds from the collected three-dimensional point cloud data of the vehicle's surrounding environment;

[0074] A clustering module, used to perform clustering processing on the ground point cloud to generate a plurality of point clusters;

[0075] The calculation module is used to extract the spatial boundary parameters of each point cluster and calculate the length and width of the rectangular area corresponding to each point cluster;

[0076] A second screening module is configured to screen out, based on the vehicle size, the rectangular areas whose length and width are both larger than the corresponding dimensions of the vehicle as candidate parking areas;

[0077] The control module is configured to execute parking path planning and automatic parking operations based on a target parking area selected by a user from the candidate parking areas.

[0078] Corresponding to the automatic parking method described in the first embodiment of the present invention, the third embodiment of the present invention further provides an automatic parking device, including:

[0079] one or more processors;

[0080] Memory;

[0081] One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, and the one or more applications are configured to execute the automatic parking method described in the aforementioned embodiment 1 of the present invention.

[0082] Corresponding to the automatic parking method described in the aforementioned embodiment 1 of the present invention, embodiment 4 of the present invention further provides a computer program product, including computer instructions, which instruct a computer device to perform operations corresponding to the automatic parking method described in the aforementioned embodiment 1 of the present invention.

[0083] Preferably, the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may be any conventional processor. The processor is the control center of the device, and various parts of the device are connected using various interfaces and lines.

[0084] The memory mainly includes a program storage area and a data storage area, wherein the program storage area can store an operating system, an application program required for at least one function, etc., and the data storage area can store related data, etc. In addition, the memory can be a high-speed random access memory, or a non-volatile memory, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, and a flash card, etc., or the memory can also be other volatile solid-state storage devices.

[0085] It should be noted that the above-mentioned device may include but is not limited to a processor and a memory, which can be understood by those skilled in the art.

[0086] For the working principle and process of the above embodiment, please refer to the description of the above embodiment of the present invention, which will not be repeated here.

[0087] As can be seen from the above description, compared with the existing technology, the present invention offers the following advantages: Addressing the inherent drawback of traditional automated parking systems, which rely on structured parking space features, the present invention innovatively introduces a quantitative assessment mechanism for three-dimensional terrain flatness, effectively addressing the technical challenges of parking area detection and vehicle mechanical protection in unstructured scenarios. By integrating highly layered filtering of LiDAR point clouds with a dual-constraint clustering algorithm, the system accurately identifies parking areas where surface roughness meets the suspension system's safety threshold, overcoming the limitations of traditional solutions that rely solely on two-dimensional plane detection based on parking space lines or obstacle spacing. In non-standard scenarios such as outdoor camping and temporary parking lots, the present invention enables flat area detection with adaptive matching of vehicle length and width, extending parking operations to complex terrain such as grass and sand, and improving its applicability in unstructured scenarios compared to existing technologies. Furthermore, the selection of parking areas based on three-dimensional terrain features significantly reduces the stress concentration effects caused by long-term parking on sloped or uneven surfaces, reduces the risk of frame torsional deformation, and extends the service life of the suspension system, providing technical support for reliable parking of smart vehicles in specialized scenarios such as wilderness exploration and emergency rescue.

[0088] The above disclosure is merely a preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. An automatic parking method, characterized in that: The following steps are involved: Filtering ground point clouds from the collected three-dimensional point cloud data of the vehicle's surrounding environment; performing clustering processing on the ground point cloud to generate a plurality of point clusters; Extract the spatial boundary parameters of each point cluster and calculate the length and width of the rectangular area corresponding to each point cluster; Based on the vehicle size, the rectangular area having a length and a width both larger than the corresponding size of the vehicle is selected as a candidate parking area; Based on the target parking area selected by the user from the candidate parking areas, parking path planning and automatic parking operations are performed.

2. The method according to claim 1, characterized in that The clustering process of the ground point cloud to generate a plurality of point clusters specifically includes: The filtered ground point clouds are spatially traversed, and adjacent point clouds that simultaneously meet the following conditions are clustered to obtain several point clusters: The Euclidean distance in the horizontal projection plane is ≤ the preset distance threshold; The height difference in the z-axis direction is less than or equal to the preset height difference threshold; The vehicle's center of mass is taken as the origin, and the z-axis is perpendicular to the horizontal plane.

3. The method according to claim 2, characterized in that The step of extracting the spatial boundary parameters of each point cluster and calculating the length and width of the rectangular area corresponding to each point cluster specifically includes: Traverse the x and y coordinate values ​​of all points in each point cluster and determine the maximum x-axis coordinate x max , the minimum x-axis coordinate x min , the maximum value of the y-axis coordinate y max , minimum y-axis coordinate y min ; Calculate the extreme value spacing on the horizontal projection plane to obtain the rectangular area size of each point cluster, where the length L = x max -x min , width W = y max -y min .

4. The method according to claim 3, characterized in that The selecting, based on the vehicle size, the rectangular area having a length and a width both larger than the corresponding size of the vehicle as a candidate parking area specifically includes: The length L and width W of the rectangular area corresponding to each point cluster are respectively compared with the vehicle's own length L veh , width W veh Compare and select the ones that satisfy L>L veh And W>W veh The rectangular area is taken as the candidate parking area.

5. The method according to claim 4, characterized in that The performing of parking path planning and automatic parking operation based on the target parking area selected by the user from the candidate parking areas specifically includes: Outputting the candidate parking area to an onboard human-computer interaction interface; Based on a target parking area selected by a user from the candidate parking areas, outputting the coordinates of four corner points of the target parking area, and performing route planning using an automatic parking planning algorithm; The vehicle is controlled to park in the target parking area according to the planned route.

6. The method according to claim 1, characterized in that The step of filtering out the ground point cloud from the collected three-dimensional point cloud data of the vehicle's surrounding environment specifically includes: Performing point cloud hierarchical filtering based on a height empirical threshold H on the collected three-dimensional point cloud data of the vehicle's surrounding environment; Filtering out the point cloud whose z-axis coordinate satisfies z∈[-H-δ,-H+δ] as the ground point cloud; The three-dimensional point cloud data adopts the vehicle center coordinate system, with the positive x-axis pointing in the direction of vehicle travel and the origin located at the vehicle center of mass; the positive y-axis extends perpendicular to the x-axis to the right, and the positive z-axis extends perpendicular to the horizontal plane and upward; δ is the height tolerance parameter.

7. The method according to claim 6, characterized in that Also includes: Combined with the reflection intensity value in the three-dimensional point cloud data, the ground point cloud is verified twice to remove point cloud data of non-ground materials.

8. An automatic parking device, characterized in that: include: A first screening module is used to screen out ground point clouds from the collected three-dimensional point cloud data of the vehicle's surrounding environment; A clustering module, used to perform clustering processing on the ground point cloud to generate a plurality of point clusters; The calculation module is used to extract the spatial boundary parameters of each point cluster and calculate the length and width of the rectangular area corresponding to each point cluster; A second screening module is configured to screen out, based on the vehicle size, the rectangular areas whose length and width are both larger than the corresponding dimensions of the vehicle as candidate parking areas; The control module is configured to execute parking path planning and automatic parking operations based on a target parking area selected by a user from the candidate parking areas.

9. An automatic parking device, characterized in that: include: one or more processors; Memory; One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, and the one or more applications are configured to execute the automatic parking method according to any one of claims 1 to 7.

10. A computer program product, characterized in that The method comprises computer instructions, wherein the computer instructions instruct a computer device to perform operations corresponding to the method according to any one of claims 1 to 7.

Images