Unmanned excavator construction environment point cloud filtering method and system

Abstract

The invention discloses a point cloud filtering method and system for a construction environment of an unmanned excavator. The point cloud filtering method for the construction environment of the unmanned excavator comprises the following steps: acquiring a point cloud of a global construction environment before construction; generating prior terrain information based on the point cloud of the global construction environment before construction; and obtaining a point cloud of a global construction environment in a construction process, and filtering the point cloud of the global construction environment in the construction process based on the prior terrain information. The method is suitable for the smooth material distribution condition, good in filtering effect, high in accuracy and short in consumed time.

Description

Technical Field

[0001] The present invention belongs to the technical field of point cloud filtering, and in particular relates to a point cloud filtering method and system for an unmanned excavator construction environment. Background Art

[0002] The safe and precise construction of unmanned excavators requires a perception system with precise perception. LiDAR is an important component of the perception system of unmanned excavators. It is usually installed above the excavator cab. Its main function is to scan the construction environment of the excavator and generate corresponding point clouds, providing accurate operation data for the decision-making and planning system of the unmanned excavator. During the construction process of the unmanned excavator, the dry working surface will generate a lot of dust. The LiDAR will scan the dust and generate corresponding dust noise. Dust noise will interfere with the decision-making and planning of the excavator and needs to be eliminated.

[0003] In 2020, Park et al. (Park Ji Il, Park Jihyuk, and Kim Kyung Soo. "Fast and Accurate Desnowing Algorithm for LiDAR Point Clouds." IEEE ACCESS 8. (2020): 160202-160212.) proposed the "Fast and Accurate Desnowing Algorithm for Lidar Point Clouds". Based on the fact that the intensity of snow point clouds is lower than that of surrounding objects, point clouds with low intensity are eliminated; at the same time, the radius filtering method is used to calculate the number of point clouds within the preset radius, and point clouds with low intensity but not snow are retained. This scheme mainly relies on the intensity attribute of noise points for noise filtering, but in actual industrial control, there are many objects with similar intensity to noise points, which can easily cause incorrect filtering.

[0004] In 2022, Geng Yangyang et al. (Geng Yangyang. Road detection based on LiDAR in complex weather. 2022. Xi'an University of Technology, MA thesis. doi: 10.27391 / d.cnki.gxagu.2022.000179.) proposed "Road inspection based on LiDAR in complex weather", using the statistical outlier filtering (SOR) method to filter out LiDAR noise in rainy and snowy days. This solution is mainly suitable for filtering low-density, small amounts of noise. For dense dust generated by non-institutional construction environments such as mining areas, the statistical outlier filtering (SOR) method cannot solve it.

[0005] In the Chinese invention patent application with publication number CN116106933A, a method for sensing harsh environments based on multi-lidar fusion is disclosed. The method obtains ground point cloud and non-ground point cloud based on original point cloud data, constructs point cloud map and performs PCA analysis to obtain the fused global map table, which can realize the recognition of various harsh environments such as dust, rain, snow, etc. This solution uses three-dimensional grids, which consumes a lot of resources and takes a long time in specific calculations.

[0006] In the Chinese invention patent application with publication number CN112083441A, an obstacle detection method based on deep fusion of laser radar and millimeter wave radar is disclosed. The original point cloud of laser radar and millimeter wave radar are preprocessed, and joint calibration is performed between the two to establish a depth perception model and a rain, fog and dust perception model to eliminate the interference of rain, fog and dust. This solution uses two sensors, laser radar and millimeter wave radar. The sensor cost is high, and the two sensors need to be jointly calibrated, which is complicated to operate and has errors. Summary of the invention

[0007] In view of the above problems, the present invention proposes a point cloud filtering method and system for an unmanned excavator construction environment, which is suitable for working conditions with smooth material distribution, has good filtering effect, high accuracy and short time consumption.

[0008] In order to achieve the above technical objectives and the above technical effects, the present invention is implemented through the following technical solutions: In a first aspect, the present invention provides a point cloud filtering method for an unmanned excavator construction environment, comprising: Obtaining a point cloud of the global construction environment before construction; Generate prior terrain information based on a point cloud of the global construction environment before the construction; A point cloud of a global construction environment during the construction process is obtained, and the point cloud of the global construction environment during the construction process is filtered based on the prior terrain information.

[0009] In combination with the first aspect, optionally, a method for acquiring a point cloud of a global construction environment includes: Acquire a first construction environment point cloud collected by using a laser radar, wherein the laser radar is disposed on an upper vehicle of an unmanned excavator; De-noising the first construction environment point cloud to generate a second construction environment point cloud from which the excavator's own point cloud is removed; Based on the rotational transformation relationship and the translational transformation relationship between the laser radar coordinate system and the vehicle coordinate system, and the rotational transformation relationship and the translational transformation relationship between the vehicle coordinate system and the vehicle coordinate system, the second construction environment point cloud is transformed from the laser radar coordinate system to the vehicle coordinate system to generate a third construction environment point cloud; All the third construction environment point clouds during the laser radar's rotation around the vehicle are accumulated frame by frame to generate a point cloud of the global construction environment.

[0010] In combination with the first aspect, optionally, the method for acquiring the point cloud of the global construction environment further includes: Eliminate the point cloud that exceeds the construction area in the third construction environment point cloud to generate a fourth construction environment point cloud corresponding to the construction area; All the fourth construction environment point clouds during the laser radar's rotation around the vehicle are accumulated frame by frame to generate a point cloud of the global construction environment.

[0011] In combination with the first aspect, optionally, the method for acquiring the point cloud of the global construction environment further includes: Eliminating dust and noise points in the fourth construction environment point cloud to generate a fifth construction environment point cloud; All the fifth construction environment point clouds during the laser radar's rotation around the vehicle are accumulated frame by frame to generate a point cloud of the global construction environment.

[0012] In combination with the first aspect, optionally, the method for generating the priori terrain information includes: Generate a two-dimensional grid, the two-dimensional grid is in the positive quadrant of the coordinate system X and Y of the vehicle, the grid is recorded as grid[i][j], [i][j] is the subscript index, i is the index in the X direction; j is the index in the Y direction; The x value and y value of the global construction environment point cloud before the construction are increased by offset_x and offset_y, respectively, where offset_x and offset_y are preset offsets, and assigned to the corresponding grid; Calculating the characteristic value of each grid based on the z value of the point cloud of the global construction environment; Based on the feature value of each grid, prior terrain information is generated.

[0013] In combination with the first aspect, optionally, a method for calculating the eigenvalue of each grid includes: If the number of point clouds in a grid is greater than 1, the eigenvalue of the grid is the maximum z value of all point clouds in the grid; If there is no point cloud in the grid, the eigenvalues ​​of the grid's neighboring grids are used as the eigenvalues ​​of the grid.

[0014] In combination with the first aspect, optionally, filtering the point cloud of the global construction environment during the construction process based on the prior terrain information includes: The x value and y value of the global construction environment point cloud of the construction process are increased by offset_x and offset_y respectively, and attributed to the corresponding grid; According to the prior terrain information, the characteristic value of each grid is obtained; Compare the point cloud in the grid with the eigenvalue of the grid. If the z value of the point cloud is greater than the eigenvalue of the grid, the point is a dust noise point cloud. The point cloud is removed to complete the filtering.

[0015] In combination with the first aspect, optionally, when attributing the global construction environment point cloud to a corresponding grid, the following steps are included: Any point cloud in the global construction environment point cloud is recorded as point_g (point_g.x, point_g.y), where point_g.x is the x value of the point cloud and point_g.y is the y value of the point cloud; The X-direction index is calculated by [point_g.x / x_index], and the Y-direction index is calculated by [point_g.y / y_index], where x_index is the preset step size in the X direction and y_index is the preset step size in the Y direction; Point clouds with the same index are assigned to the same mesh.

[0016] In a second aspect, the present invention provides a point cloud filtering system for an unmanned excavator construction environment, comprising a connected laser radar and a controller; The laser radar is arranged on the upper vehicle of the unmanned excavator; The controller comprises: Excavator working status publishing module, used to issue rotation instructions and prior terrain information extraction instructions; A global point cloud acquisition module of the construction environment is connected to the excavator working status publishing module. When the global point cloud acquisition module of the construction environment receives a rotation instruction, it acquires a point cloud of the global construction environment before construction. A priori terrain information extraction module is connected to the excavator working state publishing module and the global point cloud acquisition module of the construction environment respectively. When the global point cloud acquisition module of the construction environment receives the priori terrain information extraction instruction, it generates priori terrain information based on the point cloud of the global construction environment before the construction. The noise filtering module is respectively connected to the global point cloud acquisition module and the prior terrain information extraction module of the construction environment, and is used to obtain the point cloud of the global construction environment during the construction process from the global point cloud acquisition module of the construction environment, and filter the point cloud of the global construction environment during the construction process based on the prior terrain information obtained from the prior terrain information extraction module.

[0017] In a third aspect, the present invention provides a point cloud filtering system for an unmanned excavator construction environment, including a storage medium and a processor; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the method according to any one of the first aspects.

[0018] Compared with the prior art, the present invention has the following beneficial effects: The invention proposes a point cloud filtering method and system for an unmanned excavator construction environment, which is suitable for working conditions where materials are smoothly distributed, and has good filtering effect, high accuracy and short time consumption.

[0019] The present invention proposes to project the global point cloud of the construction environment before construction onto a two-dimensional plane, and perform gridding processing, and use the maximum z value of the point cloud in the grid as the eigenvalue to form prior terrain information. The prior terrain information can be generated quickly and effectively, laying a foundation for noise filtering of the global point cloud of the construction environment in the subsequent construction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative labor, among which: Figure 1 A schematic diagram of the structure of a point cloud filtering system for an unmanned excavator construction environment according to an embodiment of the present invention; Figure 2 A schematic diagram of the principle of a point cloud filtering system for an unmanned excavator construction environment according to an embodiment of the present invention; Figure 3 A flowchart of a method for generating a point cloud of a global construction environment according to an embodiment of the present invention; Figure 4 A flow chart of a method for generating prior terrain information according to an embodiment of the present invention; Figure 5 A flowchart of filtering a point cloud of a global construction environment according to an embodiment of the present invention; Figure 6 A two-dimensional grid diagram of an embodiment of the present invention; Figure 7 A schematic diagram of a point cloud after being assigned to a two-dimensional grid according to an embodiment of the present invention; in: 1-Radar laser; 2-Get on the vehicle; 3-Get off the vehicle; 4-Rotary encoder; 5-Material; 6-Dust. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention. Example 1

[0023] An embodiment of the present invention provides a method for filtering a point cloud of an unmanned excavator construction environment, comprising the following steps: (1) Obtaining a point cloud of the global construction environment before construction; (2) generating prior terrain information based on a point cloud of the global construction environment before the construction; (3) Obtaining a point cloud of the global construction environment during the construction process, and filtering the point cloud of the global construction environment during the construction process based on the prior terrain information.

[0024] In a specific implementation of the embodiment of the present invention, a method for acquiring a point cloud of a global construction environment includes: Acquire a first construction environment point cloud collected by using a laser radar 1, wherein the laser radar 1 is provided on a vehicle 2 of an unmanned excavator; Denoising the first construction environment point cloud to generate a second construction environment point cloud from which the excavator's own point cloud is excluded; in a specific implementation process, the second construction environment point cloud is generated by filtering the excavator's own point cloud in the vehicle coordinate system using the method proposed in the Chinese invention patent application with publication number CN118823123A, and eliminating the excavator's own point clouds such as the boom, arm and bucket, to obtain the second construction environment point cloud from which the excavator's own point cloud is excluded; Based on the rotation transformation relationship (i.e., rotation matrix) and translation transformation relationship (i.e., translation matrix) between the laser radar coordinate system and the vehicle coordinate system, and the rotation transformation relationship (i.e., rotation matrix) and translation transformation relationship (i.e., translation matrix) between the vehicle coordinate system and the vehicle coordinate system, the second construction environment point cloud is transformed from the laser radar coordinate system to the vehicle coordinate system to generate a third construction environment point cloud; All the third construction environment point clouds during the laser radar 1 rotating around the lower vehicle 3 are accumulated frame by frame to generate a point cloud of the global construction environment.

[0025] In a specific implementation of the embodiment of the present invention, in order to improve the filtering effect, the method for acquiring the point cloud of the global construction environment further includes: Eliminate the point clouds that are beyond the construction area (i.e., the range is far away) in the third construction environment point cloud, and generate a fourth construction environment point cloud corresponding to the construction area; in the specific implementation process, the method for generating the fourth construction environment point cloud includes: in the off-vehicle coordinate system, using the CropBox filtering method, eliminating the point clouds with a far range (non-construction area), and obtaining the construction area point cloud after range filtering; All fourth construction environment point clouds during the laser radar 1 rotating around the vehicle 3 are accumulated frame by frame to generate a point cloud of the global construction environment.

[0026] When the excavator is not performing excavation, there may be a small amount of low-density dust noise. The intensity (i.e., reflection intensity) of these noise points is small and exists in a specific range. There will be changes in different scenes, different material materials, and different radars. To this end, in a specific implementation of an embodiment of the present invention, the method for obtaining the point cloud of the global construction environment also includes: Remove dust noise points in the fourth construction environment point cloud to generate a fifth construction environment point cloud; in the specific implementation process, the method for generating the fifth construction environment point cloud includes: taking ouster os0 radar as an example, in dry and homogeneous earthwork material 5, the intensity of the noise points in this case is between (0, 200). Remove the noise points with intensity between (0, 200); All the fifth construction environment point clouds during the process of the laser radar 1 rotating around the lower vehicle 3 are accumulated frame by frame to generate a point cloud of the global construction environment.

[0027] In a specific implementation of the embodiment of the present invention, the method for generating priori terrain information includes: Generate a two-dimensional grid, the step length in the X direction is recorded as x_segment_index, the step length in the Y direction is recorded as y_segment_index; the number of grids in the X direction is recorded as x_segment_number, and the number of grids in the Y direction is recorded as y_segment_number; the two-dimensional grid is in the positive quadrant of the X and Y coordinate systems of the vehicle, such as Figure 6 As shown, the grid is recorded as grid[i][j], [i][j] is the subscript index, i is the index in the X direction, and the value range is [0, x_segment_number -1]; j is the index in the Y direction, and the value range is [0, y_segment_number -1]; The x value and y value of the global construction environment point cloud before construction are increased by offset_x and offset_y, respectively, where offset_x and offset_y are offsets, both of which are preset constants to ensure conversion to a two-dimensional grid area, such as Figure 7 As shown, and belong to the corresponding grid; Calculating the characteristic value of each grid based on the z value of the point cloud of the global construction environment; Based on the feature value of each grid, prior terrain information is generated.

[0028] In a specific implementation of the embodiment of the present invention, a method for calculating the eigenvalue of each grid includes: If the number of point clouds in a grid is greater than 1, the eigenvalue of the grid is the maximum z value of all point clouds in the grid; If there is no point cloud in the grid, the eigenvalues ​​of the grid's neighboring grids are used as the eigenvalues ​​of the grid.

[0029] In a specific implementation of the embodiment of the present invention, filtering the point cloud of the global construction environment during the construction process based on the prior terrain information includes: The x value and y value of the global construction environment point cloud of the construction process are increased by offset_x and offset_y respectively, and attributed to the corresponding grid; According to the prior terrain information, the characteristic value of each grid is obtained; Compare the point cloud in the grid with the eigenvalue of the grid. If the z value of the point cloud is greater than the eigenvalue of the grid, the point is a dust noise point cloud. The point cloud is removed to complete the filtering.

[0030] In a specific implementation of the embodiment of the present invention, when attributing the global construction environment point cloud to the corresponding grid, the following steps are included: Any point cloud in the global construction environment point cloud is recorded as point_g (point_g.x, point_g.y), where point_g.x is the x value of the point cloud and point_g.y is the y value of the point cloud; The X-direction index is calculated by [point_g.x / x_index], and the Y-direction index is calculated by [point_g.y / y_index], where x_index is the preset step size in the X direction, y_index is the preset step size in the Y direction, and [] indicates rounding; Point clouds with the same index are grouped into the same grid. Example 2

[0031] The present invention provides a point cloud filtering system for an unmanned excavator construction environment, comprising a connected laser radar 1 and a controller; The laser radar 1 is arranged on the upper vehicle 2 of the unmanned excavator; The controller comprises: Excavator working status publishing module, used to issue rotation instructions and prior terrain information extraction instructions; A global point cloud acquisition module of the construction environment is connected to the excavator working status publishing module. When the global point cloud acquisition module of the construction environment receives a rotation instruction, it acquires a point cloud of the global construction environment before construction. A priori terrain information extraction module is connected to the excavator working state publishing module and the global point cloud acquisition module of the construction environment respectively. When the global point cloud acquisition module of the construction environment receives the priori terrain information extraction instruction, it generates priori terrain information based on the point cloud of the global construction environment before the construction. The noise filtering module is respectively connected to the global point cloud acquisition module and the prior terrain information extraction module of the construction environment, and is used to obtain the point cloud of the global construction environment during the construction process from the global point cloud acquisition module of the construction environment, and filter the point cloud of the global construction environment during the construction process based on the prior terrain information obtained from the prior terrain information extraction module.

[0032] The working process of the point cloud filtering system for the construction environment of an unmanned excavator in an embodiment of the present invention is described in detail below in conjunction with a specific implementation manner.

[0033] The unmanned excavator construction environment point cloud filtering system provided by the present invention is suitable for the excavator high-platform digging scene. The shape of the material 5 is a relatively gentle curved surface, and there is no vertical material. Dust 6 will be generated during the excavation process; Figure 1As shown, it includes a laser radar 1 and a controller; the laser radar 1 is arranged on the upper vehicle 2 of the unmanned excavator, and its specific position can be installed above the cab. The upper vehicle 2 is connected to the lower vehicle 3, and can rotate 360 ​​degrees around the lower vehicle 3. The rotary encoder 4 is installed on the excavator turntable, and can output the rotation angle of the upper vehicle 2 relative to the lower vehicle 3 in real time. The laser radar coordinate system, the coordinate origin is at point L2; the upper vehicle coordinate system, the coordinate origin is at point L1; there is a rotation and translation transformation relationship between the laser radar coordinate system and the upper vehicle coordinate system, R1 is the rotation matrix, and T1 is the translation matrix. The lower vehicle coordinate system, the coordinate origin is at point L0; there is a rotation and translation transformation relationship between the upper vehicle coordinate system and the lower vehicle coordinate system, R2 is the rotation matrix, calculated by the rotary encoder 4; T2 is the translation matrix.

[0034] The controller includes an excavator working status publishing module, a global point cloud acquisition module of the construction environment, a priori terrain information extraction module and a noise filtering module. Figure 2 As shown; Among them, Figure 3 As shown, the global point cloud acquisition module of the construction environment performs the following steps: 1) Transform the point cloud from the radar coordinate system to the vehicle coordinate system: Using the rotation and translation transformation relationship between the laser radar coordinate system and the vehicle coordinate system (R1 is the rotation matrix, T1 is the translation matrix), the construction environment point cloud is transformed from the laser radar coordinate system to the vehicle coordinate system; 2) Filtering of the excavator’s own point cloud: In the vehicle coordinate system, the method disclosed in publication number CN118823123A is used to filter the excavator's own point cloud, remove the excavator's own point cloud such as the boom, dipper and bucket, and obtain the construction environment point cloud excluding the excavator itself.

[0035] 3) Transform the point cloud from the vehicle coordinate system to the vehicle coordinate system: The point cloud of the construction environment is transformed from the on-vehicle coordinate system to the off-vehicle coordinate system by using the rotation and translation transformation relationship between the on-vehicle coordinate system and the off-vehicle coordinate system (R2 is the rotation matrix and T2 is the translation matrix).

[0036] 4) Range filtering: In the vehicle disembarkation coordinate system, the CropBox filtering method is used to eliminate point clouds with a far range (non-construction area) to obtain the point cloud of the construction area after range filtering.

[0037] 5) Small amount, low density noise filtering: When the excavator is not performing excavation, there may be a small amount of low-density dust noise. The intensity of these noise points is small and exists in a specific range. Different scenes, different materials and different radars will have different results. Taking the ouster os0 radar as an example, in dry and homogeneous earth materials, the intensity of the noise points in this case is between (0, 200). Remove the noise points with intensity between (0, 200).

[0038] 6) Point cloud accumulation: When the laser radar 1 rotates around the lower vehicle 3 with the upper vehicle 2, a small amount of point cloud filtered with low-density noise is accumulated in each frame to generate a global construction environment point cloud.

[0039] like Figure 4 As shown, the prior terrain information extraction module performs the following steps: 1) Receiving a priori terrain information extraction instruction: receiving a priori terrain information extraction instruction issued by the excavator working status release module; 2) Create a two-dimensional grid: The step length in the X direction is recorded as x_segment_index, and the step length in the Y direction is recorded as y_segment_index; The number of grids in the X direction is recorded as x_segment_number, and the number of grids in the Y direction is recorded as y_segment_number; The two-dimensional grid is in the positive quadrant of the X and Y coordinate system of the vehicle, such as Figure 6 As shown. The grid is denoted as grid[i][j], where [i][j] is the subscript index; i is the index in the X direction, ranging from [0, x_segment_number -1]; j is the index in the Y direction, ranging from [0, y_segment_number -1].

[0040] 3) Global construction environment point cloud conversion to two-dimensional grid area: In the coordinate system of getting off the vehicle, the global construction environment point cloud will have negative numbers in X and Y. Increase the X and Y of the global construction environment point cloud by offset_x and offset_y respectively to ensure conversion to the two-dimensional grid area. Figure 7 shown.

[0041] 4) Global construction environment point cloud is assigned to the grid: The global construction environment point cloud, any point cloud is denoted as point_g (point_g.x, point_g.y), point_g.x is the x value of the point cloud, point_g.y is the y value of the point cloud; the X-direction index is calculated by [point_g.x / x_index], and the Y-direction index is calculated by [point_g.y / y_index]; point clouds with the same index belong to the same grid.

[0042] 5) Generation of prior terrain information: Calculate the eigenvalue of each grid, that is, the maximum z value of all point clouds in the grid; specifically: If the number of point clouds in the grid is greater than 1, the eigenvalue of the grid is the maximum z value of the point cloud; if there is no point cloud in the grid, the eigenvalue of the grid is assigned the eigenvalue of the neighboring grid.

[0043] The stored eigenvalues ​​of all grids constitute the prior terrain information.

[0044] like Figure 5 As shown, the noise filtering module performs the following steps: A point cloud of the global construction environment during the construction process is obtained from the global point cloud acquisition module of the construction environment, and the point cloud of the global construction environment during the construction process is filtered based on the prior terrain information obtained from the prior terrain information extraction module.

[0045] The method for obtaining the point cloud of the global construction environment during the construction process includes: 1) Transform the point cloud from the radar coordinate system to the vehicle 2 coordinate system: Using the rotation and translation transformation relationship between the laser radar 1 coordinate system and the vehicle 2 coordinate system (R1 is the rotation matrix, T is the translation matrix), the construction environment point cloud is transformed from the laser radar 1 coordinate system to the vehicle 2 coordinate system; 2) Filtering of the excavator’s own point cloud: In the vehicle 2 coordinate system, the method with patent number CN118823123A is used to filter the excavator's own point cloud, eliminate the excavator's own point cloud such as boom, dipper and bucket, and obtain the construction environment point cloud excluding the excavator itself.

[0046] 3) The point cloud is transformed from the coordinate system of the vehicle 2 to the coordinate system of the vehicle 3: By using the rotation and translation transformation relationship between the coordinate system of the vehicle 2 and the coordinate system of the vehicle 3 (R2 is the rotation matrix and T2 is the translation matrix), the construction environment point cloud is transformed from the coordinate system of the vehicle 2 to the coordinate system of the vehicle 3.

[0047] 4) Range filtering: In the coordinate system of the vehicle 3, the CropBox filtering method is used to eliminate the point clouds with a long range (non-construction area) to obtain the point cloud of the construction area after range filtering.

[0048] During the filtering process, the following steps are performed: 5) Point cloud belongs to the grid: A two-dimensional grid established by a module for extracting prior terrain information; After the range filtering, any point cloud in the construction area is recorded as point_p, where point_p.x is the x value of the point cloud and point_p.y is the y value of the point cloud. The X-direction index is calculated by point_p.x / x_index, and the Y-direction index is calculated by point_p.y / y_index. Point clouds with the same index are assigned to the same grid.

[0049] 6) Determination of eigenvalues ​​within the grid: According to the prior terrain information, the eigenvalues ​​in each grid are obtained; for the range-filtered point cloud assigned to the grid, the point cloud in the grid is compared with the eigenvalue of the grid. If the z value of the point cloud is greater than the eigenvalue of the grid, the point is a dust noise point cloud and is removed; otherwise, it is a denoised construction environment point cloud and is retained. Example 3

[0050] Based on the same inventive concept as in Example 1, an embodiment of the present invention provides a point cloud filtering system for an unmanned excavator construction environment, including a storage medium and a processor; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the method according to any one of the embodiments 1.

[0051] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0052] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0053] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0054] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0055] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

[0056] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A point cloud filtering method for unmanned excavator construction environment, characterized in that: include: Obtaining a point cloud of the global construction environment before construction; Generate prior terrain information based on a point cloud of the global construction environment before the construction; A point cloud of a global construction environment during the construction process is obtained, and the point cloud of the global construction environment during the construction process is filtered based on the prior terrain information.

2. According to claim 1, a point cloud filtering method for unmanned excavator construction environment is characterized in that: The methods for obtaining the point cloud of the global construction environment include: Acquire a first construction environment point cloud collected by using a laser radar, wherein the laser radar is disposed on an upper vehicle of an unmanned excavator; De-noising the first construction environment point cloud to generate a second construction environment point cloud from which the excavator's own point cloud is removed; Based on the rotational transformation relationship and the translational transformation relationship between the laser radar coordinate system and the vehicle coordinate system, and the rotational transformation relationship and the translational transformation relationship between the vehicle coordinate system and the vehicle coordinate system, the second construction environment point cloud is transformed from the laser radar coordinate system to the vehicle coordinate system to generate a third construction environment point cloud; All the third construction environment point clouds during the laser radar's rotation around the vehicle are accumulated frame by frame to generate a point cloud of the global construction environment.

3. The point cloud filtering method for unmanned excavator construction environment according to claim 2 is characterized in that: The method for obtaining the point cloud of the global construction environment also includes: Eliminate the point cloud that exceeds the construction area in the third construction environment point cloud to generate a fourth construction environment point cloud corresponding to the construction area; All the fourth construction environment point clouds during the laser radar's rotation around the vehicle are accumulated frame by frame to generate a point cloud of the global construction environment.

4. The point cloud filtering method for unmanned excavator construction environment according to claim 3 is characterized in that: The method for obtaining the point cloud of the global construction environment also includes: Eliminating dust and noise points in the fourth construction environment point cloud to generate a fifth construction environment point cloud; All the fifth construction environment point clouds during the laser radar's rotation around the vehicle are accumulated frame by frame to generate a point cloud of the global construction environment.

5. The point cloud filtering method for unmanned excavator construction environment according to claim 1 is characterized in that: The method for generating the priori terrain information comprises: Generate a two-dimensional grid, the two-dimensional grid is in the positive quadrant of the coordinate system X and Y of the vehicle, the grid is recorded as grid[i][j], [i][j] is the subscript index, i is the index in the X direction; j is the index in the Y direction; The x value and y value of the global construction environment point cloud before the construction are increased by offset_x and offset_y, respectively, where offset_x and offset_y are preset offsets, and assigned to the corresponding grid; Calculating the characteristic value of each grid based on the z value of the point cloud of the global construction environment; Based on the feature value of each grid, prior terrain information is generated.

6. The point cloud filtering method for unmanned excavator construction environment according to claim 5 is characterized by: The calculation method of the eigenvalue of each grid includes: If the number of point clouds in a grid is greater than 1, the eigenvalue of the grid is the maximum z value of all point clouds in the grid; If there is no point cloud in the grid, the eigenvalues ​​of the grid's neighboring grids are used as the eigenvalues ​​of the grid.

7. The point cloud filtering method for unmanned excavator construction environment according to claim 5 is characterized by: The filtering of the point cloud of the global construction environment during the construction process based on the prior terrain information includes: The x value and y value of the global construction environment point cloud of the construction process are increased by offset_x and offset_y respectively, and attributed to the corresponding grid; According to the prior terrain information, the characteristic value of each grid is obtained; Compare the point cloud in the grid with the eigenvalue of the grid. If the z value of the point cloud is greater than the eigenvalue of the grid, the point is a dust noise point cloud. The point cloud is removed to complete the filtering.

8. A point cloud filtering method for unmanned excavator construction environment according to claim 5 or 7, characterized in that: When assigning the global construction environment point cloud to the corresponding grid, the following steps are included: Any point cloud in the global construction environment point cloud is recorded as point_g (point_g.x, point_g.y), where point_g.x is the x value of the point cloud and point_g.y is the y value of the point cloud; The X-direction index is calculated by [point_g.x / x_index], and the Y-direction index is calculated by [point_g.y / y_index], where x_index is the preset step size in the X direction and y_index is the preset step size in the Y direction; Point clouds with the same index are assigned to the same mesh.

9. A point cloud filtering system for unmanned excavator construction environment, characterized in that: Includes connected LiDAR and controller; The laser radar is arranged on the upper vehicle of the unmanned excavator; The controller comprises: Excavator working status publishing module, used to issue rotation instructions and prior terrain information extraction instructions; A global point cloud acquisition module of the construction environment is connected to the excavator working status publishing module. When the global point cloud acquisition module of the construction environment receives a rotation instruction, it acquires a point cloud of the global construction environment before construction. A priori terrain information extraction module is connected to the excavator working state publishing module and the global point cloud acquisition module of the construction environment respectively. When the global point cloud acquisition module of the construction environment receives the priori terrain information extraction instruction, it generates priori terrain information based on the point cloud of the global construction environment before the construction. The noise filtering module is respectively connected to the global point cloud acquisition module and the prior terrain information extraction module of the construction environment, and is used to obtain the point cloud of the global construction environment during the construction process from the global point cloud acquisition module of the construction environment, and filter the point cloud of the global construction environment during the construction process based on the prior terrain information obtained from the prior terrain information extraction module.

10. A point cloud filtering system for unmanned excavator construction environment, characterized in that: including storage media and processors; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the method according to any one of claims 1-8.

Images