A point cloud background generation method, device and communication equipment

Abstract

The present invention provides a point cloud background generation method, device and communication equipment, and relates to the field of communication technology. The method includes: processing each frame of point cloud data of multiple frames of point cloud data to obtain initial background point cloud information and initial motion point cloud information; determining multiple voxel units to which the initial motion point cloud information belongs and recording the number of frames corresponding to each voxel unit; determining the threshold corresponding to each voxel unit based on the error information of the roadside equipment; comparing the number of frames corresponding to the voxel unit with the threshold to determine whether the point cloud information of the voxel unit belongs to the background point cloud information; if the point cloud information of the voxel unit does not belong to the background point cloud information, generating target point cloud background information according to the initial background point cloud information; if the point cloud information of the voxel unit belongs to the background point cloud information, generating target point cloud background information according to the point cloud information of the voxel unit and the initial background point cloud information. The present invention can solve the problem of background leakage in the current point cloud background generation method.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a point cloud background generation method, device and communication equipment. Background Art

[0002] There are currently two main methods for generating roadside point cloud backgrounds. One is to manually select background frames without dynamic targets through data association, and associate the data frames with the background frame space points one-to-one. Then, the points whose distance between the two points between the frames is less than the threshold are identified as static target points. The other is to voxelize the space, import multi-frame background data into voxels, aggregate and project the multi-frame background frame point clouds into voxels, check the number of points in the voxels, and determine whether it is a background voxel. If the number of points is greater than the threshold, it is considered to be the background. If it is less than the threshold, it is considered that the voxel has a dynamic target passing through it and is not considered as the background.

[0003] The background generation method based on data association requires manual collection of background frames without dynamic targets, and non-fixed background point clouds such as leaves in the background are difficult to remove through data association. In the voxelized background removal algorithm, the point cloud in the background frame is not required to be completely static. However, considering factors such as device jitter and lidar error, when the distance is far, the points of the same laser in different frame point clouds will be in different voxels, resulting in a smaller distribution of background aggregated point clouds in the farther background voxels, which are easily identified as dynamic targets, especially ground background missed detection, which is easy to affect the subsequent target recognition algorithm. Summary of the invention

[0004] The purpose of the present invention is to provide a point cloud background generation method, device and communication equipment to solve the problem of background leakage in the current point cloud background generation method.

[0005] To achieve the above object, an embodiment of the present invention provides a point cloud background generation method, comprising:

[0006] Obtain multi-frame point cloud data collected by roadside equipment;

[0007] Processing each frame of point cloud data separately to obtain initial background point cloud information and initial motion point cloud information, and for the initial motion point cloud information of each frame of point cloud data, determining a plurality of voxel units to which the initial motion point cloud information belongs, and recording the number of frames corresponding to each voxel unit;

[0008] Determining a threshold value corresponding to each voxel unit based on the error information of the roadside equipment;

[0009] Compare the number of frames corresponding to each voxel unit with the threshold to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information;

[0010] If the point cloud information corresponding to each voxel unit does not belong to the background point cloud information, the target point cloud background information is generated according to the initial background point cloud information; if the point cloud information corresponding to the target voxel unit belongs to the background point cloud information, the target point cloud background information is generated according to the point cloud information corresponding to the target voxel unit and the initial background point cloud information; wherein the target voxel unit is at least one of the multiple voxel units.

[0011] Optionally, the processing of each frame of point cloud data to obtain initial background point cloud information and initial motion point cloud information includes:

[0012] The following processing is performed on each frame of point cloud data:

[0013] Fitting point cloud information on a nearly smooth surface in the point cloud data based on an octree region growing algorithm as the initial background point cloud information;

[0014] Point cloud information other than the initial background point cloud information in the point cloud data is determined as the initial motion point cloud information.

[0015] Optionally, for each frame of the initial motion point cloud data, the multiple voxel units to which the initial motion point cloud information belongs are respectively determined, and the number of frames corresponding to each voxel unit is recorded, including:

[0016] The initial motion point cloud information of each frame of point cloud data is aggregated and projected to each voxel unit of the octree, and the counter corresponding to the voxel unit with the point cloud information projected is increased by 1.

[0017] Optionally, determining a threshold corresponding to each voxel unit based on the error information of the roadside equipment includes:

[0018] For each voxel unit, the threshold of the voxel unit is calculated according to the following steps:

[0019] Determining the number of voxel units that can be projected within an error region when the roadside equipment collects point cloud data corresponding to the voxel unit;

[0020] According to the number and the number of frames of the point cloud data, the number of frames in which the point cloud information corresponding to the voxel unit appears in the multiple frames of point cloud data is calculated;

[0021] A threshold corresponding to the voxel unit is determined according to the frame number.

[0022] Optionally, the determining the number of projectable voxel units within an error region when the roadside equipment collects point cloud data corresponding to the voxel unit includes:

[0023] Calculating the area of ​​the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit;

[0024] Calculate the orthographic projection area of ​​the voxel unit on the error region;

[0025] A ratio between the area of ​​the error region and the orthographic projection area is determined as the number of projectable voxel units in the error region.

[0026] Optionally, the calculating the area of ​​the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit includes:

[0027] Calculate the jitter radius of the roadside device when collecting point cloud data corresponding to the voxel unit according to the distance between the roadside device and the voxel unit and the jitter angle of the roadside device;

[0028] Calculating the area of ​​the error region according to the distance from the target point to the center point of the voxel unit and the jitter radius;

[0029] The target point is: a point in the point cloud data corresponding to the voxel unit collected by the roadside equipment, which is the farthest from the center point of the voxel unit.

[0030] Optionally, determining the threshold corresponding to the voxel unit according to the frame number includes:

[0031] If the difference between the frame number and the preset coefficient is less than the target threshold, determining the target threshold as the threshold corresponding to the voxel unit;

[0032] If the difference between the frame number and the preset coefficient is greater than or equal to the target threshold, the difference between the frame number and the preset coefficient is determined as the threshold corresponding to the voxel unit.

[0033] Optionally, comparing the number of frames corresponding to each voxel unit with the threshold to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information includes:

[0034] If the number of frames corresponding to the voxel unit is greater than or equal to the threshold corresponding to the voxel unit, it is determined that the point cloud information corresponding to the voxel unit belongs to the background point cloud information;

[0035] If the number of frames corresponding to the voxel unit is less than a threshold corresponding to the voxel unit, it is determined that the point cloud information corresponding to the voxel unit does not belong to the background point cloud information.

[0036] To achieve the above object, an embodiment of the present invention provides a point cloud background generating device, comprising:

[0037] An acquisition module, used to acquire multi-frame point cloud data collected by roadside equipment;

[0038] A processing module, for processing each frame of point cloud data respectively to obtain initial background point cloud information and initial motion point cloud information, and for the initial motion point cloud information of each frame of point cloud data, determining a plurality of voxel units to which the initial motion point cloud information belongs, and recording the number of frames corresponding to each voxel unit;

[0039] A determination module, used to determine a threshold value corresponding to each voxel unit based on the error information of the roadside equipment;

[0040] A comparison module, used for comparing the number of frames corresponding to each voxel unit with the threshold value, to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information;

[0041] A generation module, used for generating target point cloud background information according to the initial background point cloud information if the point cloud information corresponding to each voxel unit does not belong to the background point cloud information; if the point cloud information corresponding to the target voxel unit belongs to the background point cloud information, generating the target point cloud background information according to the point cloud information corresponding to the target voxel unit and the initial background point cloud information; wherein the target voxel unit is at least one of the multiple voxel units.

[0042] Optionally, the processing module is further used for:

[0043] The following processing is performed on each frame of point cloud data:

[0044] Fitting point cloud information on a nearly smooth surface in the point cloud data based on an octree region growing algorithm as the initial background point cloud information;

[0045] Point cloud information other than the initial background point cloud information in the point cloud data is determined as the initial motion point cloud information.

[0046] Optionally, the processing module is further used for:

[0047] The initial motion point cloud information of each frame of point cloud data is aggregated and projected to each voxel unit of the octree, and the counter corresponding to the voxel unit with the point cloud information projected is increased by 1.

[0048] Optionally, the determining module is further used for:

[0049] For each voxel unit, the threshold of the voxel unit is calculated according to the following steps:

[0050] Determine the number of voxel units that can be projected within an error region when the roadside equipment collects point cloud data corresponding to the voxel unit;

[0051] According to the number and the number of frames of the point cloud data, the number of frames in which the point cloud information corresponding to the voxel unit appears in the multiple frames of point cloud data is calculated;

[0052] A threshold corresponding to the voxel unit is determined according to the frame number.

[0053] Optionally, the processing module includes:

[0054] A first calculation unit, used to calculate the area of ​​the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit;

[0055] A second calculation unit, used for calculating the orthographic projection area of ​​the voxel unit on the error region;

[0056] The determination unit is used to determine the ratio between the area of ​​the error region and the orthographic projection area as the number of voxel units that can be projected in the error region.

[0057] Optionally, the first computing unit includes:

[0058] A first calculation subunit is used to calculate the jitter radius when the roadside device collects point cloud data corresponding to the voxel unit according to the distance between the roadside device and the voxel unit and the jitter angle of the roadside device;

[0059] A second calculation subunit, used for calculating the area of ​​the error region according to the distance from the target point to the center point of the voxel unit and the jitter radius;

[0060] The target point is: a point in the point cloud data corresponding to the voxel unit collected by the roadside equipment, which is the farthest from the center point of the voxel unit.

[0061] Optionally, the processing module includes:

[0062] A first processing unit, configured to determine the target threshold as a threshold corresponding to the voxel unit if the difference between the frame number and the preset coefficient is less than a target threshold;

[0063] The second processing unit is configured to determine the difference between the frame order and the preset coefficient as a threshold corresponding to the voxel unit if the difference between the frame order and the preset coefficient is greater than or equal to the target threshold.

[0064] Optionally, the comparison module includes:

[0065] A first comparison unit is configured to determine that the point cloud information corresponding to the voxel unit belongs to the background point cloud information if the number of frames corresponding to the voxel unit is greater than or equal to a threshold value corresponding to the voxel unit;

[0066] The second comparison unit is used to determine that the point cloud information corresponding to the voxel unit does not belong to the background point cloud information if the number of frames corresponding to the voxel unit is less than a threshold corresponding to the voxel unit.

[0067] To achieve the above object, an embodiment of the present invention provides a communication device, including: a transceiver and a processor;

[0068] The transceiver is used to: obtain multi-frame point cloud data collected by roadside equipment;

[0069] The processor is used to: process each frame of point cloud data respectively to obtain initial background point cloud information and initial motion point cloud information, and for the initial motion point cloud information of each frame of point cloud data, determine the multiple voxel units to which the initial motion point cloud information belongs, and record the frame number corresponding to each voxel unit;

[0070] Determining a threshold value corresponding to each voxel unit based on the error information of the roadside equipment;

[0071] Compare the number of frames corresponding to each voxel unit with the threshold to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information;

[0072] If the point cloud information corresponding to each voxel unit does not belong to the background point cloud information, the target point cloud background information is generated according to the initial background point cloud information; if the point cloud information corresponding to the target voxel unit belongs to the background point cloud information, the target point cloud background information is generated according to the point cloud information corresponding to the target voxel unit and the initial background point cloud information; wherein the target voxel unit is at least one of the multiple voxel units.

[0073] Optionally, the processor is further configured to:

[0074] The following processing is performed on each frame of point cloud data:

[0075] Fitting point cloud information on a nearly smooth surface in the point cloud data based on an octree region growing algorithm as the initial background point cloud information;

[0076] Point cloud information other than the initial background point cloud information in the point cloud data is determined as the initial motion point cloud information.

[0077] Optionally, the processor is further configured to:

[0078] The initial motion point cloud information of each frame of point cloud data is aggregated and projected to each voxel unit of the octree, and the counter corresponding to the voxel unit with the point cloud information projected is increased by 1.

[0079] Optionally, the processor is further configured to:

[0080] For each voxel unit, the threshold of the voxel unit is calculated according to the following steps:

[0081] Determine the number of voxel units that can be projected within an error region when the roadside equipment collects point cloud data corresponding to the voxel unit;

[0082] According to the number and the number of frames of the point cloud data, the number of frames in which the point cloud information corresponding to the voxel unit appears in the multiple frames of point cloud data is calculated;

[0083] A threshold corresponding to the voxel unit is determined according to the frame number.

[0084] Optionally, the processor is further configured to:

[0085] Calculating the area of ​​the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit;

[0086] Calculate the orthographic projection area of ​​the voxel unit on the error region;

[0087] A ratio between the area of ​​the error region and the orthographic projection area is determined as the number of projectable voxel units in the error region.

[0088] Optionally, the processor is further configured to:

[0089] Calculate the jitter radius of the roadside device when collecting point cloud data corresponding to the voxel unit according to the distance between the roadside device and the voxel unit and the jitter angle of the roadside device;

[0090] Calculating the area of ​​the error region according to the distance from the target point to the center point of the voxel unit and the jitter radius;

[0091] The target point is: a point in the point cloud data corresponding to the voxel unit collected by the roadside equipment, which is the farthest from the center point of the voxel unit.

[0092] Optionally, the processor is further configured to:

[0093] If the difference between the frame number and the preset coefficient is less than the target threshold, determining the target threshold as the threshold corresponding to the voxel unit;

[0094] If the difference between the frame number and the preset coefficient is greater than or equal to the target threshold, the difference between the frame number and the preset coefficient is determined as the threshold corresponding to the voxel unit.

[0095] Optionally, the processor is further configured to:

[0096] If the number of frames corresponding to the voxel unit is greater than or equal to the threshold corresponding to the voxel unit, it is determined that the point cloud information corresponding to the voxel unit belongs to the background point cloud information;

[0097] If the number of frames corresponding to the voxel unit is less than a threshold corresponding to the voxel unit, it is determined that the point cloud information corresponding to the voxel unit does not belong to the background point cloud information.

[0098] To achieve the above-mentioned purpose, an embodiment of the present invention provides a communication device, including a transceiver, a processor, a memory, and a program or instruction stored in the memory and executable on the processor; when the processor executes the program or instruction, the steps in the point cloud background generation method as described above are implemented.

[0099] To achieve the above objective, an embodiment of the present invention provides a readable storage medium having a program or instruction stored thereon, wherein the program or instruction, when executed by a processor, implements the steps in the point cloud background generating method as described above.

[0100] The beneficial effects of the above technical solution of the present invention are as follows:

[0101] The embodiment of the present invention can reduce the processing amount when processing point cloud information other than the initial background point cloud information, that is, the initial motion point cloud information, by preliminarily screening out the initial background point cloud information, thereby improving the processing efficiency. And by further processing the preliminarily screened initial motion point cloud information, it is determined whether the preliminarily screened initial motion point cloud information contains the background point cloud information, thereby reducing the missed detection rate of the background point cloud information. BRIEF DESCRIPTION OF THE DRAWINGS

[0102] Figure 1 is a flow chart of a method for generating a point cloud background according to an embodiment of the present invention;

[0103] Figure 2 A flow chart of determining the number of projectable voxel units in the error area of ​​a roadside device according to an embodiment of the present invention;

[0104] Figure 3 It is a schematic diagram of a roadside device and a voxel unit i in a coordinate system according to an embodiment of the present invention;

[0105] Figure 4 is a block diagram of a point cloud background generating device according to an embodiment of the present invention;

[0106] Figure 5 One of the block diagrams of the communication device according to the embodiment of the present invention;

[0107] Figure 6 This is the second block diagram of the communication device according to the embodiment of the present invention. DETAILED DESCRIPTION

[0108] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0109] It should be understood that the references to "one embodiment" or "an embodiment" throughout the specification mean that the specific features, structures, or characteristics associated with the embodiment are included in at least one embodiment of the present invention. Therefore, the references to "in one embodiment" or "in an embodiment" appearing throughout the specification do not necessarily refer to the same embodiment. In addition, these specific features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0110] In various embodiments of the present invention, it should be understood that the size of the serial numbers of the following processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0111] In the embodiments provided in the present application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined according to A. However, it should also be understood that determining B according to A does not mean determining B only according to A, and B can also be determined according to A and / or other information.

[0112] like Figure 1 As shown, a point cloud background generation method according to an embodiment of the present invention may specifically include the following steps:

[0113] Step 11: Obtain multi-frame point cloud data collected by roadside equipment.

[0114] Optionally, the roadside equipment may collect multiple frames of point cloud data through a laser radar. The number of the multiple frames of point cloud data may be determined based on actual needs, for example, 200 frames of point cloud data may be selected for processing, but the present invention is not limited thereto.

[0115] Optionally, after acquiring multiple frames of point cloud data collected by the roadside equipment, the following step 12 may be performed for each frame of point cloud data; or, after acquiring a frame of point cloud data collected by the roadside equipment, the following step 12 may be performed, that is, corresponding processing is performed after acquiring the frame point cloud data, thereby realizing the processing of multiple frames of point cloud data, and the embodiments of the present invention are not limited to this.

[0116] Step 12: Process each frame of point cloud data separately to obtain initial background point cloud information and initial motion point cloud information, and for the initial motion point cloud information of each frame of point cloud data, determine the multiple voxel units to which the initial motion point cloud information belongs, and record the frame number corresponding to each voxel unit.

[0117] Optionally, the initial background point cloud information may be points on a nearly smooth surface in the fitted point cloud data, such as the ground, the surface corresponding to the building, the lawn, etc. The nearly smooth surface mentioned here may refer to a surface whose curvature is less than a preset threshold, and the "curvature" is only used to enable those skilled in the art to know what a "nearly smooth surface" is, and does not limit the need to fit the initial background point cloud information based on the "curvature".

[0118] Optionally, the initial motion point cloud information may refer to point cloud information in the point cloud data other than the initial background point cloud information, such as vehicles, trees, etc.

[0119] In this way, by preliminarily screening out the initial background point cloud information, it can be ensured that when processing the point cloud information other than the initial background point cloud information, that is, the initial motion point cloud information, the processing amount is reduced, thereby improving the processing efficiency.

[0120] Step 13: Based on the error information of the roadside equipment, determine the threshold corresponding to each voxel unit.

[0121] Optionally, the error information of the roadside equipment may include errors caused by the jitter of the roadside equipment itself, or errors caused by the laser radar when collecting data.

[0122] Step 14: Compare the number of frames corresponding to each voxel unit with the threshold to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information.

[0123] Optionally, the above step 14 may specifically include: if the number of frames corresponding to the voxel unit is greater than or equal to the threshold corresponding to the voxel unit, then determining that the point cloud information corresponding to the voxel unit belongs to the background point cloud information; if the number of frames corresponding to the voxel unit is less than the threshold corresponding to the voxel unit, then determining that the point cloud information corresponding to the voxel unit does not belong to the background point cloud information.

[0124] Of course, in addition to the above-mentioned method of comparing the number of frames corresponding to the voxel unit with the threshold, other methods can also be used, such as: determining whether the difference between the number of frames corresponding to the voxel unit and the threshold meets the preset conditions, etc. The embodiments of the present application are not limited to this.

[0125] Step 15: If the point cloud information corresponding to each voxel unit does not belong to the background point cloud information, then the target point cloud background information is generated according to the initial background point cloud information; if the point cloud information corresponding to the target voxel unit belongs to the background point information, then the target point cloud background information is generated according to the point cloud information corresponding to the target voxel unit and the initial background point cloud information.

[0126] The target voxel unit is at least one of the multiple voxel units.

[0127] In the above scheme, by preliminarily screening out the initial background point cloud information, it is possible to ensure that when processing the point cloud information other than the initial background point cloud information, namely the initial motion point cloud information, the processing amount is reduced, thereby improving the processing efficiency. And by further processing the initial motion point cloud information obtained by the preliminary screening, it is determined whether the initial motion point cloud information obtained by the preliminary screening contains the background point cloud information, thereby reducing the missed detection rate of the background point cloud information.

[0128] Optionally, the step of processing each frame of point cloud data separately to obtain initial background point cloud information and initial motion point cloud information may specifically include:

[0129] The following processing is performed on each frame of point cloud data:

[0130] Fitting point cloud information on a nearly smooth surface in the point cloud data based on an octree region growing algorithm as the initial background point cloud information;

[0131] Point cloud information other than the initial background point cloud information in the point cloud data is determined as the initial motion point cloud information.

[0132] Specifically, a frame of point cloud data is read, and the point cloud information on the same smooth surface in the point cloud data is fitted based on the octree region growing algorithm, and the region growing algorithm can fit the surface with curvature (that is, the smooth surface can be a plane or a surface with a certain curvature), such as: point cloud information corresponding to the surface of a building, the ground, the lawn, etc. in the point cloud data. Optionally, based on the region growing algorithm and the threshold setting, it can be determined that the points within the distance range in the normal direction belong to the same smooth surface.

[0133] In this way, the octree-based region growing algorithm can identify the initial background point cloud information in the point cloud data, and then determine the point cloud information in the point cloud data other than the initial background point cloud information as the initial motion point cloud information (or called noise information), such as vehicles, trees, etc.

[0134] Optionally, when the initial background point cloud information is obtained by identifying the initial background point cloud information based on the octree region growing algorithm, the voxel information corresponding to the initial background point cloud information can be saved in the initial background recognition result. In this way, the initial background recognition result is stored in the form of an octree, which has the advantage of a small amount of data and is convenient for docking with the algorithm.

[0135] Optionally, the step of respectively determining, for each frame of initial motion point cloud information of the point cloud data, a plurality of voxel units to which the initial motion point cloud information belongs, and recording the number of frames corresponding to each voxel unit may specifically include:

[0136] The initial motion point cloud information of each frame of point cloud data is aggregated and projected to each voxel unit of the octree, and the counter corresponding to the voxel unit with the point cloud information projected is increased by 1.

[0137] Optionally, after the initial motion point cloud information of each frame of point cloud data is aggregated and projected into each voxel unit of the octree, the corresponding voxel information can be saved in the initial motion space, and in the initial motion space, the counter corresponding to the voxel unit with the point cloud information projected is accumulated. For example: in the process of aggregating the initial motion point cloud information in the current frame of point cloud data, if the initial motion point cloud information has point cloud information corresponding to the first voxel unit, then the counter corresponding to the first voxel unit is increased by 1, and so on until all point cloud data are processed, so that the number of frames corresponding to the corresponding voxel unit can be statistically obtained.

[0138] Optionally, the step of determining a threshold value corresponding to each voxel unit based on the error information of the roadside equipment may specifically include:

[0139] For each voxel unit, the threshold of the voxel unit is calculated according to the following steps:

[0140] Determining the number of voxel units that can be projected within an error region when the roadside equipment collects point cloud data corresponding to the voxel unit;

[0141] According to the number and the number of frames of the point cloud data, the number of frames in which the point cloud information corresponding to the voxel unit appears in the multiple frames of point cloud data is calculated;

[0142] A threshold corresponding to the voxel unit is determined according to the frame number.

[0143] The farther away from the roadside equipment, the lower the point density, and the greater the impact of factors such as roadside equipment jitter. The laser (such as LiDAR) will oscillate within a certain range when emitting lasers, which may cause the same laser to emit point clouds in different frames in different voxels, and then cause the counter value of the corresponding voxel unit in the initial motion space to be low, which is easily misidentified as non-background point cloud information (i.e., motion point cloud information), that is, it leads to under-recognition of background point cloud information, especially under-recognition of ground point cloud information, which will affect the target recognition accuracy.

[0144] Thus, in the embodiment of the present invention, the error area when the roadside equipment collects point cloud data caused by the vibration of the roadside equipment, recognition errors and the like is taken into consideration, and the threshold calculation of the corresponding voxel unit is performed, that is, an adaptive threshold calculation is performed for each voxel unit, thereby ensuring the recognition accuracy of the initial motion point cloud information as background point cloud information and reducing the missed detection rate of the background point cloud information.

[0145] like Figure 2As shown, the step of determining the number of projectable voxel units within the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit may specifically include:

[0146] Step 21: Calculate the area of ​​the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit.

[0147] Optionally, the above step 21 may specifically include:

[0148] Calculate the jitter radius of the roadside device when collecting point cloud data corresponding to the voxel unit according to the distance between the roadside device and the voxel unit and the jitter angle of the roadside device;

[0149] Calculating the area of ​​the error region according to the distance from the target point to the center point of the voxel unit and the jitter radius;

[0150] The target point is: a point in the point cloud data corresponding to the voxel unit collected by the roadside equipment, which is the farthest from the center point of the voxel unit.

[0151] The following combination Figure 3 For an explanation of the above method: see Figure 3 , the coordinate origin O represents the position of a laser radar in the position of the roadside equipment, and the center point of the voxel unit i ( Figure 3 The center of the solid circle in the figure), with coordinates p(x,y,z).

[0152] When the jitter angle of the laser radar is θ, the jitter radius m when the roadside equipment collects the point cloud data corresponding to the voxel unit can be calculated by the following formula (1):

[0153] m=sin(θ)*D (1)

[0154] Where D is the distance from the laser radar to the center point of voxel unit i,

[0155] Considering Figure 3 The size of the circle with a radius of m is small, and it can be approximated that the circle with a radius of m is perpendicular to the XY plane. The distance from the target point l to the center point p of the voxel unit i is It can be calculated by the following formula (2):

[0156]

[0157] Among them, the target point l is the point farthest from the center point of the voxel unit in the point cloud data corresponding to the voxel unit collected by the roadside equipment, or it can be understood as: the focus of the extended line of the laser radar to the top q of the circle with a radius of m and the ground. H is the vertical distance from the center point p to the XY plane, that is, H=z; L is the distance from the positive projection point P' of the center point p in the XY plane to the coordinate origin O, that is,

[0158] Consider that the xy plane where the drive test equipment is located is approximately parallel to the ground, and Figure 3 The area of ​​the dotted triangle is taken as the area of ​​the error region, and the area of ​​the error region area(i) can be calculated by the following formula (3):

[0159]

[0160] Step 22: Calculate the orthographic projection area of ​​the voxel unit on the error region.

[0161] Optionally, the voxel unit i is a cube with a side length of v, then the area of ​​the orthographic projection of the voxel unit i on the surface where the error region is located is: v 2 .

[0162] Step 23: Determine the ratio between the area of ​​the error region and the orthographic projection area as the number of projectable voxel units in the error region.

[0163] Optionally, the number of projectable voxel units num in the error region i , can be calculated by the following formula (4):

[0164]

[0165] Optionally, when the number of frames of multi-frame point cloud data is n, the number of frames in which the point cloud information corresponding to the voxel unit i appears in the multi-frame point cloud data is: n / num i .

[0166] Optionally, the step of determining the threshold value corresponding to the voxel unit according to the frame number may specifically include:

[0167] If the difference between the frame number and the preset coefficient is less than the target threshold, determining the target threshold as the threshold corresponding to the voxel unit;

[0168] If the difference between the frame number and the preset coefficient is greater than or equal to the target threshold, the difference between the frame number and the preset coefficient is determined as the threshold corresponding to the voxel unit.

[0169] Optionally, the target threshold and the preset coefficient can be set according to actual needs, or according to experimental measurement results. The preset system is to better distinguish dynamic point cloud information from background point cloud information.

[0170] For example, in the embodiment of the present invention, the target threshold can be set to 5, and the threshold threshold corresponding to the voxel unit i can be expressed by the following formula (5):

[0171]

[0172] Among them, th is the preset coefficient.

[0173] In this way, in the voxelized background removal algorithm, the above-mentioned embodiment of the present invention adaptively calculates the thresholds of different voxel units based on factors such as roadside equipment jitter and lidar errors, so as to reduce the misjudgment of moving point clouds in different voxels caused by points collected by the same laser in different frames of point cloud data when the distance between the object to be collected and the roadside equipment is far, thereby reducing the missed detection rate of the ground point cloud when the background point cloud is generated, and reducing the impact of the jitter or measurement error of the roadside equipment on the subsequent roadside equipment's perception of the target correspondence.

[0174] like Figure 4 As shown, an embodiment of the present invention provides a point cloud background generating device 400, comprising:

[0175] An acquisition module 410 is used to acquire multi-frame point cloud data collected by roadside equipment;

[0176] The processing module 420 is used to process each frame of point cloud data to obtain initial background point cloud information and initial motion point cloud information, and to determine the multiple voxel units to which the initial motion point cloud information belongs for the initial motion point cloud information of each frame of point cloud data, and to record the frame number corresponding to each voxel unit;

[0177] A determination module 430, configured to determine a threshold value corresponding to each voxel unit based on the error information of the roadside equipment;

[0178] A comparison module 440 is used to compare the number of frames corresponding to each voxel unit with the threshold value to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information;

[0179] A generation module 450 is used to generate target point cloud background information based on the initial background point cloud information if the point cloud information corresponding to each voxel unit does not belong to the background point cloud information; if the point cloud information corresponding to the target voxel unit belongs to the background point cloud information, generate the target point cloud background information based on the point cloud information corresponding to the target voxel unit and the initial background point cloud information; wherein the target voxel unit is at least one of the multiple voxel units.

[0180] Optionally, the processing module 420 is further configured to:

[0181] The following processing is performed on each frame of point cloud data:

[0182] Fitting point cloud information on a nearly smooth surface in the point cloud data based on an octree region growing algorithm as the initial background point cloud information;

[0183] Point cloud information other than the initial background point cloud information in the point cloud data is determined as the initial motion point cloud information.

[0184] Optionally, the processing module 420 is further configured to:

[0185] The initial motion point cloud information of each frame of point cloud data is aggregated and projected to each voxel unit of the octree, and the counter corresponding to the voxel unit with the point cloud information projected is increased by 1.

[0186] Optionally, the determining module 430 is further configured to:

[0187] For each voxel unit, the threshold of the voxel unit is calculated according to the following steps:

[0188] Determining the number of voxel units that can be projected within an error region when the roadside equipment collects point cloud data corresponding to the voxel unit;

[0189] According to the number and the number of frames of the point cloud data, the number of frames in which the point cloud information corresponding to the voxel unit appears in the multiple frames of point cloud data is calculated;

[0190] A threshold corresponding to the voxel unit is determined according to the frame number.

[0191] Optionally, the processing module 430 includes:

[0192] A first calculation unit, used to calculate the area of ​​the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit;

[0193] A second calculation unit, used for calculating the orthographic projection area of ​​the voxel unit on the error region;

[0194] The determination unit is used to determine the ratio between the area of ​​the error region and the orthographic projection area as the number of voxel units that can be projected in the error region.

[0195] Optionally, the first computing unit includes:

[0196] A first calculation subunit is used to calculate the jitter radius when the roadside device collects point cloud data corresponding to the voxel unit according to the distance between the roadside device and the voxel unit and the jitter angle of the roadside device;

[0197] A second calculation subunit, used for calculating the area of ​​the error region according to the distance from the target point to the center point of the voxel unit and the jitter radius;

[0198] The target point is: a point in the point cloud data corresponding to the voxel unit collected by the roadside equipment, which is the farthest from the center point of the voxel unit.

[0199] Optionally, the processing module 430 includes:

[0200] A first processing unit, configured to determine the target threshold as a threshold corresponding to the voxel unit if the difference between the frame number and the preset coefficient is less than a target threshold;

[0201] The second processing unit is configured to determine the difference between the frame order and the preset coefficient as a threshold corresponding to the voxel unit if the difference between the frame order and the preset coefficient is greater than or equal to the target threshold.

[0202] Optionally, the comparison module 440 includes:

[0203] A first comparison unit is configured to determine that the point cloud information corresponding to the voxel unit belongs to the background point cloud information if the number of frames corresponding to the voxel unit is greater than or equal to a threshold value corresponding to the voxel unit;

[0204] The second comparison unit is used to determine that the point cloud information corresponding to the voxel unit does not belong to the background point cloud information if the number of frames corresponding to the voxel unit is less than a threshold corresponding to the voxel unit.

[0205] The device 400 in the embodiment of the present invention can implement each process of the above-mentioned point cloud background generation method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described here.

[0206] The device 400 in the embodiment of the present invention can reduce the processing amount when processing the point cloud information other than the initial background point cloud information, i.e., the initial motion point cloud information, by preliminarily screening out the initial background point cloud information, thereby improving the processing efficiency. And by further processing the initial motion point cloud information obtained by the preliminarily screening, it is determined whether the initial motion point cloud information obtained by the preliminarily screening contains the background point cloud information, thereby reducing the missed detection rate of the background point cloud information.

[0207] like Figure 5As shown, an embodiment of the present invention provides a communication device, including: a transceiver 520 and a processor 510;

[0208] The transceiver 520 is used to: obtain multi-frame point cloud data collected by roadside equipment;

[0209] The processor 510 is used to: process each frame of point cloud data respectively to obtain initial background point cloud information and initial motion point cloud information, and for the initial motion point cloud information of each frame of point cloud data, determine the multiple voxel units to which the initial motion point cloud information belongs, and record the frame number corresponding to each voxel unit;

[0210] Determining a threshold value corresponding to each voxel unit based on the error information of the roadside equipment;

[0211] Compare the number of frames corresponding to each voxel unit with the threshold to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information;

[0212] If the point cloud information corresponding to each voxel unit does not belong to the background point cloud information, the target point cloud background information is generated according to the initial background point cloud information; if the point cloud information corresponding to the target voxel unit belongs to the background point cloud information, the target point cloud background information is generated according to the point cloud information corresponding to the target voxel unit and the initial background point cloud information; wherein the target voxel unit is at least one of the multiple voxel units.

[0213] Optionally, the processor 510 is further configured to:

[0214] The following processing is performed on each frame of point cloud data:

[0215] Fitting point cloud information on a nearly smooth surface in the point cloud data based on an octree region growing algorithm as the initial background point cloud information;

[0216] Point cloud information other than the initial background point cloud information in the point cloud data is determined as the initial motion point cloud information.

[0217] Optionally, the processor 510 is further configured to:

[0218] The initial motion point cloud information of each frame of point cloud data is aggregated and projected to each voxel unit of the octree, and the counter corresponding to the voxel unit with the point cloud information projected is increased by 1.

[0219] Optionally, the processor 510 is further configured to:

[0220] For each voxel unit, the threshold of the voxel unit is calculated according to the following steps:

[0221] Determine the number of voxel units that can be projected within an error region when the roadside equipment collects point cloud data corresponding to the voxel unit;

[0222] According to the number and the number of frames of the point cloud data, the number of frames in which the point cloud information corresponding to the voxel unit appears in the multiple frames of point cloud data is calculated;

[0223] A threshold corresponding to the voxel unit is determined according to the frame number.

[0224] Optionally, the processor 510 is further configured to:

[0225] Calculating the area of ​​the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit;

[0226] Calculate the orthographic projection area of ​​the voxel unit on the error region;

[0227] A ratio between the area of ​​the error region and the orthographic projection area is determined as the number of projectable voxel units in the error region.

[0228] Optionally, the processor 510 is further configured to:

[0229] Calculate the jitter radius of the roadside device when collecting point cloud data corresponding to the voxel unit according to the distance between the roadside device and the voxel unit and the jitter angle of the roadside device;

[0230] Calculating the area of ​​the error region according to the distance from the target point to the center point of the voxel unit and the jitter radius;

[0231] The target point is: a point in the point cloud data corresponding to the voxel unit collected by the roadside equipment, which is the farthest from the center point of the voxel unit.

[0232] Optionally, the processor 510 is further configured to:

[0233] If the difference between the frame number and the preset coefficient is less than the target threshold, determining the target threshold as the threshold corresponding to the voxel unit;

[0234] If the difference between the frame number and the preset coefficient is greater than or equal to the target threshold, the difference between the frame number and the preset coefficient is determined as the threshold corresponding to the voxel unit.

[0235] Optionally, the processor 510 is further configured to:

[0236] If the number of frames corresponding to the voxel unit is greater than or equal to the threshold corresponding to the voxel unit, it is determined that the point cloud information corresponding to the voxel unit belongs to the background point cloud information;

[0237] If the number of frames corresponding to the voxel unit is less than a threshold corresponding to the voxel unit, it is determined that the point cloud information corresponding to the voxel unit does not belong to the background point cloud information.

[0238] The communication device 500 in the embodiment of the present invention can implement each process of the above-mentioned point cloud background generation method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described here.

[0239] The communication device 500 in the embodiment of the present invention can reduce the processing amount when processing the point cloud information other than the initial background point cloud information, i.e., the initial motion point cloud information, by preliminarily screening out the initial background point cloud information, thereby improving the processing efficiency. And by further processing the preliminarily screened initial motion point cloud information, it is determined whether the preliminarily screened initial motion point cloud information contains the background point cloud information, thereby reducing the missed detection rate of the background point cloud information.

[0240] A communication device according to another embodiment of the present invention, such as Figure 6 As shown, it includes a transceiver 610, a processor 600, a memory 620, and a program or instruction stored in the memory 620 and executable on the processor 600; when the processor 600 executes the program or instruction, the steps applied to the point cloud background generation method are implemented, which will not be repeated here to avoid repetition.

[0241] The transceiver 610 is used to receive and send data under the control of the processor 600 .

[0242] Among them, Figure 6 In the embodiment, the bus architecture may include any number of interconnected buses and bridges, specifically linking together various circuits of one or more processors represented by processor 600 and memory represented by memory 620. The bus architecture may also link together various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface. The transceiver 610 may be a plurality of components, i.e., including a transmitter and a receiver, providing a unit for communicating with various other devices on a transmission medium. The processor 600 is responsible for managing the bus architecture and general processing, and the memory 620 may store data used by the processor 600 when performing operations.

[0243] A readable storage medium according to an embodiment of the present invention stores a program or instruction thereon. When the program or instruction is executed by a processor, the steps in the point cloud background generation method as described above are implemented and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0244] The processor is a processor in the communication device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0245] It should be further explained that the terminals described in this specification include but are not limited to smart phones, tablet computers, etc., and many of the functional components described are called modules in order to more particularly emphasize the independence of their implementation methods.

[0246] In the embodiment of the present invention, module can be implemented with software so that it can be executed by various types of processors. For example, an executable code module of an identification can include one or more physical or logical blocks of computer instructions, for example, it can be constructed as an object, process or function. Nevertheless, the executable code of the identified module does not need to be physically located together, but can include different instructions stored in different positions, and when these instructions are logically combined together, it constitutes a module and realizes the specified purpose of the module.

[0247] In fact, executable code module can be a single instruction or many instructions, and can even be distributed on a plurality of different code segments, distributed among different programs, and distributed across a plurality of memory devices. Similarly, operating data can be identified in the module, and can be implemented and organized in the data structure of any appropriate type according to any appropriate form. The operating data can be collected as a single data set, or can be distributed in different locations (including on different storage devices), and can only be present on a system or network as an electronic signal at least in part.

[0248] When a module can be implemented by software, considering the level of existing hardware technology, a person skilled in the art can build a corresponding hardware circuit to implement the corresponding function of the module that can be implemented by software without considering the cost. The hardware circuit includes a conventional very large scale integration (VLSI) circuit or gate array and existing semiconductors such as logic chips, transistors, or other discrete components. The module can also be implemented by a programmable hardware device, such as a field programmable gate array, a programmable array logic, a programmable logic device, etc.

[0249] The above exemplary embodiments are described with reference to the accompanying drawings, and many different forms and embodiments are feasible without departing from the spirit and teachings of the present invention. Therefore, the present invention should not be constructed as a limitation of the exemplary embodiments proposed herein. More specifically, these exemplary embodiments are provided so that the present invention will be perfect and complete, and the scope of the present invention will be conveyed to those who are familiar with the technology. In these figures, the component sizes and relative sizes may be exaggerated for clarity. The terms used here are only based on the purpose of describing specific exemplary embodiments and are not intended to be limiting. As used herein, unless the text clearly indicates otherwise, the singular forms "one", "an" and "the" are intended to include these multiple forms. It will be further understood that the terms "including" and / or "comprising" when used in this specification indicate the presence of the features, integers, steps, operations, components and / or components, but do not exclude the presence or increase of one or more other features, integers, steps, operations, components, components and / or their groups. Unless otherwise indicated, when stated, a range of values ​​includes the upper and lower limits of that range and any subranges therebetween.

[0250] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A point cloud background generation method, characterized in that: include: Obtain multi-frame point cloud data collected by roadside equipment; Processing each frame of point cloud data separately to obtain initial background point cloud information and initial motion point cloud information, and for the initial motion point cloud information of each frame of point cloud data, determining a plurality of voxel units to which the initial motion point cloud information belongs, and recording the number of frames corresponding to each voxel unit; Determining a threshold value corresponding to each voxel unit based on the error information of the roadside equipment; Compare the number of frames corresponding to each voxel unit with the threshold to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information; If the point cloud information corresponding to each voxel unit does not belong to the background point cloud information, generating the target point cloud background information according to the initial background point cloud information; If the point cloud information corresponding to the target voxel unit belongs to the background point cloud information, the target point cloud background information is generated according to the point cloud information corresponding to the target voxel unit and the initial background point cloud information; wherein the target voxel unit is at least one of the multiple voxel units.

2. The point cloud background generation method according to claim 1, characterized in that: The point cloud data of each frame is processed separately to obtain initial background point cloud information and initial motion point cloud information, including: The following processing is performed on each frame of point cloud data: Fitting point cloud information on a nearly smooth surface in the point cloud data based on an octree region growing algorithm as the initial background point cloud information; wherein the nearly smooth surface refers to a surface whose curvature is less than a preset threshold; Point cloud information other than the initial background point cloud information in the point cloud data is determined as the initial motion point cloud information.

3. The point cloud background generation method according to claim 1, characterized in that: The initial motion point cloud information of each frame of point cloud data is respectively determined to which multiple voxel units the initial motion point cloud information belongs, and the frame number corresponding to each voxel unit is recorded, including: The initial motion point cloud information of each frame of point cloud data is aggregated and projected to each voxel unit of the octree, and the counter corresponding to the voxel unit with the point cloud information projected is increased by 1.

4. The point cloud background generation method according to claim 1, characterized in that: The step of determining a threshold value corresponding to each voxel unit based on the error information of the roadside equipment includes: For each voxel unit, the threshold of the voxel unit is calculated according to the following steps: Determine the number of voxel units that can be projected within an error region when the roadside equipment collects point cloud data corresponding to the voxel unit; According to the number and the number of frames of the point cloud data, the number of frames in which the point cloud information corresponding to the voxel unit appears in the multiple frames of point cloud data is calculated; A threshold corresponding to the voxel unit is determined according to the frame number.

5. The point cloud background generation method according to claim 4, characterized in that: The determining of the number of projectable voxel units within an error region when the roadside equipment collects point cloud data corresponding to the voxel unit comprises: Calculating the area of ​​the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit; Calculate the orthographic projection area of ​​the voxel unit on the error region; A ratio between the area of ​​the error region and the orthographic projection area is determined as the number of projectable voxel units in the error region.

6. The point cloud background generation method according to claim 5, characterized in that: The calculating the area of ​​the error region when the roadside equipment collects the point cloud data corresponding to the voxel unit includes: Calculate the jitter radius of the roadside device when collecting point cloud data corresponding to the voxel unit according to the distance between the roadside device and the voxel unit and the jitter angle of the roadside device; Calculating the area of ​​the error region according to the distance from the target point to the center point of the voxel unit and the jitter radius; The target point is: a point in the point cloud data corresponding to the voxel unit collected by the roadside equipment, which is the farthest from the center point of the voxel unit.

7. The point cloud background generation method according to claim 4, characterized in that: The step of determining the threshold value corresponding to the voxel unit according to the frame number includes: If the difference between the frame number and the preset coefficient is less than the target threshold, determining the target threshold as the threshold corresponding to the voxel unit; If the difference between the frame number and the preset coefficient is greater than or equal to the target threshold, the difference between the frame number and the preset coefficient is determined as the threshold corresponding to the voxel unit.

8. The point cloud background generation method according to claim 1, characterized in that: The step of comparing the number of frames corresponding to each voxel unit with the threshold to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information includes: If the number of frames corresponding to the voxel unit is greater than or equal to the threshold corresponding to the voxel unit, it is determined that the point cloud information corresponding to the voxel unit belongs to the background point cloud information; If the number of frames corresponding to the voxel unit is less than a threshold corresponding to the voxel unit, it is determined that the point cloud information corresponding to the voxel unit does not belong to the background point cloud information.

9. A point cloud background generating device, characterized in that: include: An acquisition module, used to acquire multi-frame point cloud data collected by roadside equipment; A processing module, for processing each frame of point cloud data respectively to obtain initial background point cloud information and initial motion point cloud information, and for the initial motion point cloud information of each frame of point cloud data, determining a plurality of voxel units to which the initial motion point cloud information belongs, and recording the number of frames corresponding to each voxel unit; A determination module, used to determine a threshold value corresponding to each voxel unit based on the error information of the roadside equipment; A comparison module, used for comparing the number of frames corresponding to each voxel unit with the threshold value, to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information; A generating module, configured to generate target point cloud background information according to the initial background point cloud information if the point cloud information corresponding to each voxel unit does not belong to the background point cloud information; If the point cloud information corresponding to the target voxel unit belongs to the background point cloud information, the target point cloud background information is generated according to the point cloud information corresponding to the target voxel unit and the initial background point cloud information; wherein the target voxel unit is at least one of the multiple voxel units.

10. A communication device, characterized in that: include: transceivers and processors; The transceiver is used to: obtain multi-frame point cloud data collected by roadside equipment; The processor is used to: process each frame of point cloud data respectively to obtain initial background point cloud information and initial motion point cloud information, and for the initial motion point cloud information of each frame of point cloud data, determine the multiple voxel units to which the initial motion point cloud information belongs, and record the frame number corresponding to each voxel unit; Determining a threshold value corresponding to each voxel unit based on the error information of the roadside equipment; Compare the number of frames corresponding to each voxel unit with the threshold to determine whether the point cloud information corresponding to the voxel unit belongs to the background point cloud information; If the point cloud information corresponding to each voxel unit does not belong to the background point cloud information, generating the target point cloud background information according to the initial background point cloud information; If the point cloud information corresponding to the target voxel unit belongs to the background point cloud information, the target point cloud background information is generated according to the point cloud information corresponding to the target voxel unit and the initial background point cloud information; wherein the target voxel unit is at least one of the multiple voxel units.

11. A communication device, comprising: A transceiver, a processor, a memory, and a program or instruction stored in the memory and executable on the processor; characterized in that when the processor executes the program or instruction, the steps of the point cloud background generation method as described in any one of claims 1 to 8 are implemented.

12. A readable storage medium having a program or instruction stored thereon, characterized in that: When the program or instruction is executed by a processor, the steps of the point cloud background generation method according to any one of claims 1 to 8 are implemented.

Images