Point cloud ground segmentation method based on port scene ground characteristics and related equipment

Abstract

The application provides a point cloud ground segmentation method based on port scene ground characteristics and related equipment. The method comprises: acquiring laser point cloud data collected by a laser sensor; dividing the laser point cloud data into a plurality of grids in the horizontal direction; performing a ground laser point cloud data identification step on the laser point cloud data in each grid, the ground laser point cloud data identification step comprising: plane fitting on the laser point cloud data in the grid; in response to the laser point cloud data in the grid being fitted into a plane and meeting ground characteristics, determining the laser point cloud data fitted into the plane in the grid as ground laser point cloud data; in response to the laser point cloud data in the grid not being fitted into a plane or the laser point cloud data in the grid not meeting ground characteristics, adjusting the size of the grid, and performing the ground laser point cloud data identification step on the laser point cloud data in the grid again. The application improves the accuracy of point cloud ground segmentation in a port scene.

Description

Technical Field

[0001] This invention relates to the field of lidar data processing, and in particular to a point cloud ground segmentation method and related equipment based on the ground characteristics of a port scene. Background Technology

[0002] LiDAR does not distinguish the surrounding ground or other obstacles; instead, it outputs them all as points in three-dimensional space. Therefore, algorithms are needed to differentiate them so that the backend can make the correct decisions.

[0003] Currently, there are several methods for ground segmentation:

[0004] The first method combines the scanning characteristics of the lidar transmitter to encode the point cloud and segment the ground based on the differences in features between adjacent points. For example, by traversing each point of each beam of a mechanical lidar, if a significant height difference is found between neighboring points, it is considered an obstacle rather than the ground. However, this method only works for mechanically scanning lidars and requires specific installation angles. For solid-state lidars, or lidars that need to be vertically mounted for scanning, the recognition effect is poor.

[0005] The second method is the ray method. The surrounding space is divided into several sector-shaped regions along the angle and meridian. The center of each sector is the lidar. The ground, from the side view of each sector, forms a ray extending from the lidar to infinity. Along the ray direction, from near to far, each point is compared to the previous ground point to determine if it meets the slope and height difference thresholds. However, this method suffers from a chain reaction; if the previous point is incorrectly identified, subsequent points will also be incorrect. Furthermore, this process must be sequential and cannot be parallelized, resulting in significant time consumption.

[0006] The third approach involves fusion with other sensors, such as cameras. This involves semantic segmentation from the camera data followed by fusion with the point cloud to achieve ground segmentation. However, this method relies on other sensors and requires significant computational power for algorithm development and system performance.

[0007] In addition, the above methods have several problems: 1) They only consider the features of points, without extracting the overall features of all points globally or locally, or the extracted features are singular and not flexible enough; 2) The laser calibration results have a significant impact on the ground recognition effect. If the calibration error is large, or if the laser is jolted during movement, the probability of segmentation failure will be greatly increased; 3) They are all based on a single laser, but in practical applications, especially in the field of autonomous driving, the use of multiple lasers has become the norm. Multiple lasers have mutual calibration errors. When multiple lasers are stitched together, different lasers seeing the same ground may produce multiple layers of ground results, causing ground segmentation based on stitched light borrowing to fail.

[0008] Therefore, how to reduce the impact of laser calibration on ground segmentation, improve the flexibility of feature extraction, and enhance the accuracy of laser point cloud ground segmentation are technical problems that urgently need to be solved by those skilled in the art. Summary of the Invention

[0009] In order to overcome the shortcomings of the prior art, this invention provides a point cloud ground segmentation method, device, electronic device, and storage medium based on the ground characteristics of port scenes, so as to reduce the impact of laser calibration on ground segmentation and improve the flexibility of feature extraction and the accuracy of laser point cloud ground segmentation.

[0010] According to one aspect of the present invention, a point cloud ground segmentation method based on ground characteristics of a port scene is provided, comprising:

[0011] Acquire laser point cloud data collected by the laser sensor;

[0012] In the horizontal direction, the laser point cloud data is divided into multiple grids;

[0013] For each grid cell containing laser point cloud data, a ground laser point cloud data identification step is performed, which includes:

[0014] Perform planar fitting on the laser point cloud data in the grid;

[0015] In response to the laser point cloud data in the grid being fitted to a plane and conforming to ground features, the laser point cloud data in the grid that is fitted to a plane is determined as ground laser point cloud data;

[0016] In response to the laser point cloud data in the grid not being fitted to a plane, or the laser point cloud data in the grid not conforming to ground features, the size of the grid is adjusted, and the ground laser point cloud data recognition step is performed again on the laser point cloud data in the grid.

[0017] In some embodiments of this application, in response to the laser point cloud data in the grid not being fitted to a plane, or the laser point cloud data in the grid not conforming to ground features, the method further includes:

[0018] Query the recognition results of the ground laser point cloud data recognition steps performed on adjacent grid cells;

[0019] In response to the adjacent grid having ground laser point cloud data, the laser point cloud data in the grid that is located on the same plane as the ground laser point cloud data of the adjacent grid is determined as quasi-ground laser point cloud data.

[0020] In some embodiments of this application, in response to the grid having terrestrial laser point cloud data and quasi-terrestrial laser point cloud data, the terrestrial laser point cloud data identification step further includes:

[0021] The ground laser point cloud data is adjusted based on the quasi-ground laser point cloud data.

[0022] In some embodiments of this application, the multiple grids are of the same size.

[0023] In some embodiments of this application, dividing the laser point cloud data into multiple grids in the horizontal direction includes:

[0024] The initial size of the grid is determined based on the vehicle width and / or road width of the vehicle where the laser sensor is located.

[0025] In some embodiments of this application, adjusting the size of the grid includes:

[0026] The size of the grid is reduced to one-quarter of the original grid size.

[0027] In some embodiments of this application, it also includes:

[0028] Acquire ground laser point cloud data collected by multiple laser sensors;

[0029] In the common field of view of multiple laser sensors, the ground laser point cloud data of the multiple laser sensors are aligned to perform error calibration on the laser point cloud data collected by the multiple laser sensors.

[0030] According to another aspect of this application, a point cloud ground segmentation device based on the ground characteristics of a port scene is also provided, comprising:

[0031] The laser point cloud acquisition module is used to acquire laser point cloud data collected by the laser sensor.

[0032] The grid division module is used to divide the laser point cloud data into multiple grids in the horizontal direction;

[0033] The ground identification module is used to perform a ground laser point cloud data identification step on the laser point cloud data in each grid cell. The ground laser point cloud data identification step includes:

[0034] Perform planar fitting on the laser point cloud data in the grid;

[0035] In response to the laser point cloud data in the grid being fitted to a plane and conforming to ground features, the laser point cloud data in the grid that is fitted to a plane is determined as ground laser point cloud data;

[0036] In response to the laser point cloud data in the grid not being fitted to a plane, or the laser point cloud data in the grid not conforming to ground features, the size of the grid is adjusted, and the ground laser point cloud data recognition step is performed again on the laser point cloud data in the grid.

[0037] According to another aspect of the present invention, an electronic device is also provided, the electronic device comprising: a processor; and a storage medium having a computer program stored thereon, the computer program being executed by the processor to perform the steps described above.

[0038] According to another aspect of the present invention, a storage medium is also provided, on which a computer program is stored, the computer program being executed by a processor to perform the steps described above.

[0039] Compared with the prior art, the advantages of this invention are:

[0040] The laser point cloud data is divided into multiple grids in the horizontal direction. Plane fitting is performed on the laser point cloud data in each grid. When the laser point cloud data in a grid fits a plane and conforms to ground features, the laser point cloud data in that grid that fits a plane is identified as ground laser point cloud data. When the laser point cloud data in a grid does not fit a plane, or does not conform to ground features, the size of the grid is adjusted, and the ground laser point cloud data identification step is performed again on the laser point cloud data in that grid. Therefore, the gridding method reduces the impact of laser calibration on ground segmentation and improves the flexibility of feature extraction and the accuracy of laser point cloud ground segmentation. This is particularly suitable for port scenarios where the ground is uneven or has multiple planes, curved surfaces, etc. (such as ramps or potholes), ensuring the accuracy of uneven ground segmentation. Attached Figure Description

[0041] The above and other features and advantages of the present invention will become more apparent from a detailed description of exemplary embodiments thereof with reference to the accompanying drawings.

[0042] Figure 1 A flowchart of a point cloud ground segmentation method based on ground characteristics of a port scene according to an embodiment of the present invention is shown;

[0043] Figure 2 A flowchart of a point cloud ground segmentation method based on ground characteristics of a port scene according to a specific embodiment of the present invention is shown;

[0044] Figure 3 A schematic diagram of the initial grid division according to an embodiment of the present invention is shown;

[0045] Figure 4A schematic diagram of grid size adjustment according to an embodiment of the present invention is shown;

[0046] Figure 5 A flowchart illustrating the determination of ground laser point cloud data based on adjacent grids according to an embodiment of the present invention is shown;

[0047] Figure 6 A block diagram of a point cloud ground segmentation device based on the ground characteristics of a port scene according to an embodiment of the present invention is shown;

[0048] Figure 7 This illustration schematically depicts a computer-readable storage medium according to an exemplary embodiment of the present disclosure.

[0049] Figure 8 The schematic diagram illustrates an electronic device according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0050] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this disclosure will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0051] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0052] To address the shortcomings of existing technologies, this invention provides a point cloud ground segmentation method based on the ground characteristics of a port scene. See below. Figure 1 , Figure 1 A flowchart of a point cloud ground segmentation method based on the ground characteristics of a port scene according to an embodiment of the present invention is shown. Figure 1 The steps are as follows:

[0053] Step S110: Acquire laser point cloud data collected by the laser sensor;

[0054] Step S120: Divide the laser point cloud data into multiple grids in the horizontal direction;

[0055] Step S130: Perform a ground laser point cloud data identification step on the laser point cloud data in each grid cell, the ground laser point cloud data identification step including:

[0056] Perform planar fitting on the laser point cloud data in the grid;

[0057] In response to the laser point cloud data in the grid being fitted to a plane and conforming to ground features, the laser point cloud data in the grid that is fitted to a plane is determined as ground laser point cloud data;

[0058] In response to the laser point cloud data in the grid not being fitted to a plane, or the laser point cloud data in the grid not conforming to ground features, the size of the grid is adjusted, and the ground laser point cloud data recognition step is performed again on the laser point cloud data in the grid.

[0059] In the point cloud ground segmentation method based on port scene ground characteristics provided by this invention, the laser point cloud data is divided into multiple grids in the horizontal direction. Plane fitting is performed on the laser point cloud data in each grid. When the laser point cloud data in the grid fits a plane and conforms to ground features, the laser point cloud data fitted to the plane in the grid is determined as ground laser point cloud data. When the laser point cloud data in the grid does not fit a plane, or does not conform to ground features, the size of the grid is adjusted, and the ground laser point cloud data recognition step is performed again on the laser point cloud data in the grid. Therefore, by using a gridded approach, the influence of laser calibration on ground segmentation is reduced, and the flexibility of feature extraction and the accuracy of laser point cloud ground segmentation are improved. This method is particularly suitable for port scenes where the ground is uneven or has multiple planes, curved surfaces, etc. (such as ramps or potholes), ensuring the accuracy of uneven ground segmentation.

[0060] See below. Figure 2 , Figure 2 A flowchart of a point cloud ground segmentation method based on the ground characteristics of a port scene according to a specific embodiment of the present invention is shown. Figure 2 The following steps are shown:

[0061] Step S201: Acquire laser point cloud data collected by the laser sensor.

[0062] Step S202: Divide the laser point cloud data into multiple grids in the horizontal direction.

[0063] Specifically, the horizontal plane is parallel to the ground. Step S202 divides the laser point cloud data into multiple grids according to the initial grid size. The multiple grids obtained in step S202 have the same size. Further, the initial size of the grid can be determined based on the vehicle width and / or road width of the vehicle where the laser sensor is located. For example, if the vehicle width is 3 meters and the road width is greater than 3 meters, the initial size of the grid can be determined as 6*6 meters (i.e., twice the vehicle width as the width and height of the grid). This allows for the determination of the initial grid size based on the vehicle width and / or road width in the actual application scenario. By limiting the initial size, each grid is closer to the actual scenario, and the number of initial grids can be controlled, improving the efficiency of subsequent grid processing.

[0064] The grid division can be referenced. Figure 3 , Figure 3 A schematic diagram of the initial grid division according to an embodiment of the present invention is shown. Grids 301 to 309 can be obtained based on the initial dimensions. Figure 3 This is merely illustrative; the number of grid cells is not limited by this.

[0065] Step S203: Select one grid cell as the grid cell to be processed.

[0066] Step S204: Perform plane fitting on the laser point cloud data in the grid to be processed.

[0067] Specifically, principal component analysis, random consistency algorithm, etc. can be used to perform plane fitting of the laser point cloud in the grid to be processed. This application can achieve more variations, which will not be elaborated here.

[0068] Step S205: Determine whether the laser point cloud data in the grid to be processed is fitted to a plane and conforms to ground features.

[0069] Specifically, ground features may include, but are not limited to, the lowest plane height among multiple fitted planes, plane height differences within a set ground height difference threshold range, and the normal vector of the fitted plane being nearly perpendicular to the horizontal plane. This application may set various methods for judging ground features, which will not be elaborated here.

[0070] If step S205 determines that it is yes, then step S207 is executed: the laser point cloud data fitted to the plane in the grid is determined as ground laser point cloud data.

[0071] If step S205 determines no, then step S206 is executed: adjust the grid size and divide the grid to be processed.

[0072] Specifically, the size of the grid can be reduced to one-quarter of the original grid size. Taking an initial grid size of 6*6 meters as an example, in step S206, the grid size can be adjusted to 3*3 meters, and the grid to be processed can be divided into 3*3 meter grids. After step S206, step S203 is executed again to select one grid from the four divided grids as the grid to be processed, until the four grids have been traversed through steps S204 to S207.

[0073] Combination Figure 3 and Figure 4 Plane fitting and ground feature matching can be performed sequentially on the laser point cloud data within grids 301 to 309. The laser point cloud data within grids 301 to 304 can all undergo plane fitting and ground feature matching to obtain the ground laser point cloud data for grids 301 to 304. The laser point cloud data within grid 305 does not fit to a plane, so grid 305 is divided into grids 305A to 305D. Plane fitting and ground feature matching are then performed sequentially on the laser point cloud data within grids 305A to 305D. If plane fitting and ground feature matching can be performed on all of them, then plane fitting and ground feature matching are performed on grid 306. If the laser point cloud data within any grid from 305A to 305D does not fit to a plane, then that grid is further divided.

[0074] Furthermore, to avoid multiple iterations, the number of times step S206 adjusts the size can be set, or the minimum size adjusted in step S206 can be limited. For example, step S206 can be limited to a maximum of two executions or the minimum grid size can be 1.5*1.5 meters. Thus, when the grid size is adjusted to 1.5*1.5 meters, even if the laser point cloud data in the grid does not fit a plane, or the laser point cloud data in the grid does not conform to the ground features, the size of the grid will not be adjusted further.

[0075] Step S208: Determine whether each grid cell has been traversed.

[0076] If step S208 determines no, then step S203 is executed to take the next grid cell as the grid cell to be processed.

[0077] If step S208 is determined to be yes, then step S209 is executed: perform ground segmentation based on ground laser point cloud data.

[0078] Specifically, after ground segmentation, ground laser point cloud data can be deleted from the laser point cloud data, so that the remaining laser point cloud data can be used for obstacle matching and recognition.

[0079] exist Figure 2In the illustrated embodiment, grid division can be performed based on the fitting and matching results when performing plane fitting and ground feature matching on the laser point cloud data in each grid. In some variations, after performing plane fitting and ground feature matching on the laser point cloud data in each grid, the laser point cloud data in the grids that failed to fit / match can be adjusted based on the fitting and matching results. For example, plane fitting and ground feature matching can be performed on the laser point cloud data in grids 301 to 309. Among them, the laser point cloud data in grids 301 to 304 and 306 to 309 can all be subjected to plane fitting and ground feature matching. Then, grid division is performed on grid 305 to obtain grids 305A to 305D, and plane fitting and ground feature matching are performed on the laser point cloud data in grids 305A to 305D in sequence. This application can implement more variations, which will not be elaborated here.

[0080] See below. Figure 5 , Figure 5 A flowchart illustrating the determination of ground laser point cloud data based on adjacent grids according to an embodiment of the present invention is shown. In response to the laser point cloud data in the grid not fitting a plane, or the laser point cloud data in the grid not conforming to ground features, further actions may be performed as follows: Figure 5 The steps shown are as follows:

[0081] Step S410: Query the recognition results of the ground laser point cloud data recognition step performed on adjacent grid cells.

[0082] Step S420: In response to the adjacent grid having ground laser point cloud data, the laser point cloud data in the grid that is located on the same plane as the ground laser point cloud data of the adjacent grid is determined as quasi-ground laser point cloud data.

[0083] Specifically, Figure 5 The steps shown can be applied to graticles that, after limiting the number of graticle size adjustments / minimum graticle size, still cannot be fitted to a plane or matched with ground features (using quasi-ground laser point cloud data as the ground laser point cloud data for that graticle), can be fitted using adjacent graticles. Figure 2 The ground laser point cloud data obtained by the steps shown is supplemented to obtain complete ground laser point cloud data.

[0084] In other embodiments, steps S410 and S420 can be performed on each grid that cannot be fitted to a plane or match ground features, thereby determining the ground laser point cloud data of the grid by referring to quasi-ground laser point cloud data and the adjusted ground laser point cloud data. For example, either quasi-ground laser point cloud data or the adjusted ground laser point cloud data can be selected as the ground laser point cloud data of the grid. Alternatively, the ground laser point cloud data can be adjusted based on the quasi-ground laser point cloud data (the union of the quasi-ground laser point cloud data and the adjusted ground laser point cloud data can be used as the ground laser point cloud data of the grid).

[0085] In some specific embodiments, steps S410 to S420 can be applied to embodiments where, after traversing the laser point cloud data in each grid and performing planar fitting and ground feature matching, the laser point cloud data in grids that failed to fit / match are adjusted based on the fitting and matching results. Because the traversal is performed first, all grids that can be matched through planar fitting and ground feature matching, along with their ground laser point cloud data, can be obtained before executing steps S410 and S420. Therefore, when executing steps S410 and S420, a larger number of adjacent grids can be referenced, resulting in more accurate quasi-ground laser point cloud data.

[0086] In a further embodiment, steps S410 and S420 can be performed across grids. Figure 3 For example, when the laser point cloud data in grid 301 is fitted and matched to obtain ground laser point cloud data, and the fitting / matching of laser point cloud data in grids 304 and 307 fails, the quasi-ground laser point cloud data of grid 304 can be determined based on the ground laser point cloud data of grid 301 through steps S410 and S420; the quasi-ground laser point cloud data of grid 307 can be determined based on the (quasi)ground laser point cloud data of grid 304 through steps S410 and S420. Thus, through a recursive relationship, more quasi-ground laser point cloud data of grids can be obtained for reference and supplementation in the processing of ground laser point cloud data.

[0087] Furthermore, considering that multiple laser sensors are typically installed on a vehicle, this application can perform the following on each laser sensor: Figure 1 The steps are shown in step 2. Therefore, after acquiring ground laser point cloud data from multiple laser sensors, the ground laser point cloud data from the multiple laser sensors can be aligned within their shared field of view to perform error calibration on the laser point cloud data acquired by the multiple laser sensors. Thus, the laser point cloud data from multiple laser sensors are aligned to facilitate the stitching of each laser point cloud data set. After removing the ground laser point cloud data, accurate obstacle identification and matching can be performed.

[0088] This application fully considers potential errors in lidar calibration and issues such as vibration and jitter during movement. It also specifically addresses calibration errors between multiple lidars to ensure consistent output. Furthermore, recognizing that the surrounding ground is not perfectly flat and may contain uneven surfaces, multiple planes, or curved surfaces (e.g., slopes, potholes), a gridded approach is used to segment the ground into individual grids, achieving the desired segmentation effect. Different grid levels not only ensure accurate segmentation of uneven ground but also effectively suppress over-segmentation of small targets on the ground caused by sparse point clouds. Moreover, this application utilizes parallel programming, resulting in low system computational consumption and ensuring real-time performance on most chips used in autonomous driving. The number of algorithm iterations can be adaptively adjusted based on the chip's computing power, significantly improving the versatility of this application across different computing platforms.

[0089] The above are merely several specific implementations of the point cloud ground segmentation method based on the ground characteristics of port scenes according to the present invention. Each implementation can be implemented independently or in combination, and the present invention is not intended to limit it. Furthermore, the flowchart of the present invention is merely illustrative, and the execution order between the steps is not limited thereto. The splitting, merging, sequential exchange, and other synchronous or asynchronous execution methods of the steps are all within the protection scope of the present invention.

[0090] The present invention also provides a point cloud ground segmentation device based on the ground characteristics of a port scene. Figure 6 A block diagram of a point cloud ground segmentation device based on port scene ground characteristics according to an embodiment of the present invention is shown. The point cloud ground segmentation device 500 based on port scene ground characteristics includes a laser point cloud acquisition module 510, a grid division module 520, and a ground recognition module 530.

[0091] The laser point cloud acquisition module 510 is used to acquire laser point cloud data collected by the laser sensor;

[0092] The grid division module 520 is used to divide the laser point cloud data into multiple grids in the horizontal direction;

[0093] The ground recognition module 530 is used to perform a ground laser point cloud data recognition step on the laser point cloud data in each grid, the ground laser point cloud data recognition step including:

[0094] Perform planar fitting on the laser point cloud data in the grid;

[0095] In response to the laser point cloud data in the grid being fitted to a plane and conforming to ground features, the laser point cloud data in the grid that is fitted to a plane is determined as ground laser point cloud data;

[0096] In response to the laser point cloud data in the grid not being fitted to a plane, or the laser point cloud data in the grid not conforming to ground features, the size of the grid is adjusted, and the ground laser point cloud data recognition step is performed again on the laser point cloud data in the grid.

[0097] In the point cloud ground segmentation device based on port scene ground characteristics provided by this invention, the laser point cloud data is divided into multiple grids in the horizontal direction; the laser point cloud data in each grid is subjected to planar fitting; when the laser point cloud data in the grid fits into a plane and conforms to ground features, the laser point cloud data fitted into a plane in the grid is determined as ground laser point cloud data; when the laser point cloud data in the grid does not fit into a plane, or the laser point cloud data in the grid does not conform to ground features, the size of the grid is adjusted, and the ground laser point cloud data recognition step is performed again on the laser point cloud data in the grid. Thus, by using a gridded approach, the influence of laser calibration on ground segmentation is reduced, and the flexibility of feature extraction and the accuracy of laser point cloud ground segmentation are improved. This is particularly suitable for port scenes where the ground is uneven or has multiple planes, curved surfaces, etc. (such as ramps or potholes), ensuring the accuracy of segmenting uneven ground.

[0098] Figure 6 The illustrations of the point cloud ground segmentation device 500 based on port scene ground characteristics provided by this invention are merely schematic representations. Without departing from the inventive concept, the splitting, merging, and addition of modules are all within the scope of protection of this invention. The point cloud ground segmentation device 300 based on port scene ground characteristics provided by this invention can be implemented by software, hardware, firmware, plugins, and any combination thereof; this invention is not limited thereto.

[0099] In exemplary embodiments of this disclosure, a computer-readable storage medium is also provided, having stored thereon a computer program that, when executed by, for example, a processor, can implement the steps of the point cloud ground segmentation method based on port scene ground characteristics described in any of the above embodiments. In some possible implementations, various aspects of the invention can also be implemented as a program product comprising program code that, when run on a terminal device, causes the terminal device to perform the steps of the various exemplary embodiments of the invention described in the section on the point cloud ground segmentation method based on port scene ground characteristics described in this specification.

[0100] refer to Figure 7As shown, a program product 800 for implementing the above-described method according to an embodiment of the present invention is described. This product may employ a portable compact disc read-only memory (CD-ROM) and include program code, and may run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, the readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0101] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0102] The computer-readable storage medium may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, capable of transmitting, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0103] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the tenant's computing device, partially on the tenant's device, as a standalone software package, partially on the tenant's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the tenant's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0104] In an exemplary embodiment of this disclosure, an electronic device is also provided, which may include a processor and a memory for storing executable instructions of the processor. The processor is configured to perform the steps of the point cloud ground segmentation method based on port scene ground characteristics described in any of the above embodiments by executing the executable instructions.

[0105] Those skilled in the art will understand that various aspects of the present invention can be implemented as systems, methods, or program products. Therefore, various aspects of the present invention can be specifically implemented in the following forms: entirely hardware implementations, entirely software implementations (including firmware, microcode, etc.), or implementations combining hardware and software aspects, collectively referred to herein as “circuits,” “modules,” or “systems.”

[0106] The following reference Figure 8 To describe an electronic device 600 according to this embodiment of the present invention. Figure 8 The electronic device 600 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0107] like Figure 8 As shown, the electronic device 600 is presented in the form of a general-purpose computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different system components (including storage unit 620 and processing unit 610), a display unit 640, etc.

[0108] The storage unit stores program code that can be executed by the processing unit 610, causing the processing unit 610 to perform the steps described in the section on point cloud ground segmentation methods based on port scene ground characteristics described in this specification, according to various exemplary embodiments of the present invention. For example, the processing unit 610 can perform, as follows: Figure 1 The steps are shown in the figure.

[0109] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache storage unit 6202, and may further include a read-only memory unit (ROM) 6203.

[0110] The storage unit 620 may also include a program / utility 6204 having a set (at least one) program module 6205, such program module 6205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0111] Bus 630 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0112] Electronic device 600 can also communicate with one or more external devices 700 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable tenants to interact with electronic device 600, and / or with any device that enables electronic device 600 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 650. Furthermore, electronic device 600 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 660. Network adapter 660 can communicate with other modules of electronic device 600 via bus 630. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0113] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, or network device, etc.) to execute the above-described point cloud ground segmentation method based on port scene ground characteristics according to the embodiments of this disclosure.

[0114] Compared with the prior art, the advantages of this invention are:

[0115] The laser point cloud data is divided into multiple grids in the horizontal direction. Plane fitting is performed on the laser point cloud data in each grid. When the laser point cloud data in a grid fits a plane and conforms to ground features, the laser point cloud data in that grid that fits a plane is identified as ground laser point cloud data. When the laser point cloud data in a grid does not fit a plane, or does not conform to ground features, the size of the grid is adjusted, and the ground laser point cloud data identification step is performed again on the laser point cloud data in that grid. Therefore, the gridding method reduces the impact of laser calibration on ground segmentation and improves the flexibility of feature extraction and the accuracy of laser point cloud ground segmentation. This is particularly suitable for port scenarios where the ground is uneven or has multiple planes, curved surfaces, etc. (such as ramps or potholes), ensuring the accuracy of uneven ground segmentation.

[0116] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the appended claims.

Claims

1. A point cloud ground segmentation method based on ground characteristics of a port scene, characterized in that, include: Acquire laser point cloud data collected by the laser sensor; In the horizontal direction, the laser point cloud data is divided into multiple grids; For each grid cell containing laser point cloud data, a ground laser point cloud data identification step is performed, which includes: Perform planar fitting on the laser point cloud data in the grid; In response to the laser point cloud data in the grid being fitted to a plane and conforming to ground features, the laser point cloud data in the grid that is fitted to a plane is determined as ground laser point cloud data; In response to the laser point cloud data in the grid not being fitted to a plane, or the laser point cloud data in the grid not conforming to ground features, the identification results of the ground laser point cloud data identification step performed on adjacent grids are queried; in response to the adjacent grid having ground laser point cloud data, the laser point cloud data in the grid that is located on the same plane as the ground laser point cloud data of the adjacent grid is identified as quasi-ground laser point cloud data; the size of the grid is adjusted, and the ground laser point cloud data identification step is performed again on the laser point cloud data in the grid, and the ground laser point cloud data of the grid is determined based on the quasi-ground laser point cloud data and the adjusted ground laser point cloud data.

2. The point cloud ground segmentation method based on port scene ground characteristics as described in claim 1, characterized in that, The multiple grids are of the same size.

3. The point cloud ground segmentation method based on port scene ground characteristics as described in claim 2, characterized in that, Dividing the laser point cloud data into multiple grids in the horizontal direction includes: The initial size of the grid is determined based on the vehicle width and / or road width of the vehicle where the laser sensor is located.

4. The point cloud ground segmentation method based on port scene ground characteristics as described in claim 2, characterized in that, Adjusting the size of the grid includes: The size of the grid is reduced to one-quarter of the original grid size.

5. The point cloud ground segmentation method based on port scene ground characteristics as described in claim 1, characterized in that, Also includes: Acquire ground laser point cloud data collected by multiple laser sensors; In the common field of view of multiple laser sensors, the ground laser point cloud data of the multiple laser sensors are aligned to perform error calibration on the laser point cloud data collected by the multiple laser sensors.

6. A point cloud ground segmentation device based on the ground characteristics of a port scene, characterized in that, include: The laser point cloud acquisition module is used to acquire laser point cloud data collected by the laser sensor. The grid division module is used to divide the laser point cloud data into multiple grids in the horizontal direction; The ground identification module is used to perform a ground laser point cloud data identification step on the laser point cloud data in each grid cell. The ground laser point cloud data identification step includes: Perform planar fitting on the laser point cloud data in the grid; In response to the laser point cloud data in the grid being fitted to a plane and conforming to ground features, the laser point cloud data in the grid that is fitted to a plane is determined as ground laser point cloud data; In response to the laser point cloud data in the grid not being fitted to a plane, or the laser point cloud data in the grid not conforming to ground features, the identification results of the ground laser point cloud data identification step performed on adjacent grids are queried; in response to the adjacent grid having ground laser point cloud data, the laser point cloud data in the grid that is located on the same plane as the ground laser point cloud data of the adjacent grid is identified as quasi-ground laser point cloud data; the size of the grid is adjusted, and the ground laser point cloud data identification step is performed again on the laser point cloud data in the grid, and the ground laser point cloud data of the grid is determined based on the quasi-ground laser point cloud data and the adjusted ground laser point cloud data.

7. An electronic device, characterized in that, The electronic device includes: processor; A storage medium having a computer program stored thereon, the computer program being executed by the processor to perform the point cloud ground segmentation method based on the ground characteristics of a port scene as described in any one of claims 1 to 5.

8. A storage medium, characterized in that, The storage medium stores a computer program, which is executed by a processor to perform the point cloud ground segmentation method based on the ground characteristics of a port scene as described in any one of claims 1 to 5.

Images