A method for determining point cloud noise and an intelligent mobile device

Abstract

The present disclosure provides a method for determining point cloud noise and an intelligent mobile device. The present disclosure uses a pre-established preset point cloud map as prior information, and uses the vertical void space in the grid of the preset point cloud map to check the point data in the frame point cloud obtained by real-time scanning, so as to remove the useless points and / or the point data of noise in the frame point cloud. The present disclosure does not need to set a discrimination threshold, is not affected by environmental changes, and has a stable filtering effect. It avoids the statistics of massive data and can be completed by using grid coordinate indexing and comparison of spatial regions, reduces the amount of data processing, and improves the efficiency of identifying noise. At the same time, it does not rely on model training and online deployment, and does not rely on the support of high-performance hardware, reducing the usage cost.

Description

Technical Field

[0001] The present disclosure relates to the technical field of intelligent vehicles, and more particularly, to a method for determining point cloud noise points and an intelligent mobile device. Background Art

[0002] An intelligent vehicle is an integrated system that combines functions such as environmental perception, planning and decision-making, and multi-level assisted driving. It comprehensively applies technologies such as computers, modern sensors, information fusion, communication, artificial intelligence, and automatic control to provide safety, reliability, and comfort for intelligent vehicles. Among them, lidar combined with a high-precision point cloud map can provide good positioning results. However, in rainy, snowy, foggy, and dusty weather, lidar will generate a large number of noise points, which greatly affects the accuracy and robustness of point cloud positioning and brings great difficulties to positioning.

[0003] Currently, the method for determining point cloud noise points needs to calculate the statistical information and feature information of all points in the point cloud, which is time-consuming and requires the support of high-performance hardware, and is prone to over-removal or under-removal of noise points.

[0004] Therefore, the present disclosure provides a method for determining point cloud noise points to solve one of the above technical problems. Summary of the Invention

[0005] The purpose of the present disclosure is to provide a method for determining point cloud noise points and an intelligent mobile device, which can solve at least one of the above-mentioned technical problems. The specific solutions are as follows:

[0006] According to a specific embodiment of the present disclosure, in a first aspect, the present disclosure provides a method for determining point cloud noise points, including:

[0007] Obtain the position information and frame point cloud of the intelligent vehicle, where the frame point cloud includes a plurality of point data, and both the position information and the plurality of point data are set in a preset first coordinate system;

[0008] Intercept a local point cloud map in a preset point cloud map based on the position information, where the local point cloud map is set in a preset second coordinate system and includes a plurality of grids;

[0009] Obtain the horizontal projection area of each vertical void space in any grid projected onto a preset horizontal plane based on the local point cloud map;

[0010] Project the coordinates of each point data in the preset first coordinate system onto the local point cloud map in the preset second coordinate system with reference to the position information in the preset first coordinate system, and obtain the horizontal projection coordinates of each point data projected onto the preset horizontal plane;

[0011] When the horizontal projection coordinates of any point data exist in any horizontal projection area, it is determined that the point data corresponding to the horizontal projection coordinates is point cloud noise.

[0012] Optionally, the method further includes:

[0013] While obtaining the horizontal projection coordinates of each point data projected onto a preset horizontal plane, the grid coordinate index of each point data in a preset second coordinate system is also obtained, where the grid coordinate index refers to the coordinates of the grid on the coordinate plane formed by the first coordinate axis and the second coordinate axis in the preset second coordinate system.

[0014] Optionally, the obtaining the grid coordinate index of each point data in a preset second coordinate system includes:

[0015] Obtaining the first coordinate of any point data in the direction of the first coordinate axis, and

[0016] Obtaining the second coordinate of the point data in the direction of the second coordinate axis;

[0017] Calculating the quotient of the first coordinate and the preset length of the grid in the direction of the first coordinate axis to obtain the index first coordinate of the grid coordinate index in the direction of the first coordinate axis;

[0018] Calculating the quotient of the second coordinate and the preset width of the grid in the direction of the second coordinate axis to obtain the index second coordinate of the grid coordinate index in the direction of the second coordinate axis.

[0019] Optionally, the when the horizontal projection coordinates of any point data exist in any horizontal projection area, it is determined that the point data corresponding to the horizontal projection coordinates is point cloud noise, includes:

[0020] Obtaining all the horizontal projection areas of the grid based on the grid coordinate index of the grid to which any point data belongs;

[0021] Traversing all the horizontal projection areas of the grid based on the horizontal projection coordinates of the point data;

[0022] When the horizontal projection coordinates of the point data exist in any horizontal projection area of the grid, it is determined that the point data is point cloud noise.

[0023] Optionally, each horizontal projection area is a rectangular projection, and two adjacent sides of the rectangular projection are respectively parallel to the first coordinate axis and the second coordinate axis in the preset second coordinate system;

[0024] Correspondingly, the when the horizontal projection coordinates of the point data exist in any horizontal projection area of the grid, it is determined that the point data is point cloud noise, includes:

[0025] Obtain the minimum value and the maximum value of any horizontal projection area on the first coordinate axis, generate a first value range of the horizontal projection area on the first coordinate axis, and

[0026] obtain the minimum value and the maximum value of the horizontal projection area on the second coordinate axis, and generate a second value range of the horizontal projection area on the second coordinate axis;

[0027] When the value of the horizontal projection coordinate of the point data on the first coordinate axis is within the first value range and the value of the horizontal projection coordinate of the point data on the second coordinate axis is within the second value range, determine that the point data is point cloud noise.

[0028] Optionally, the projecting the coordinates of each point data in the preset first coordinate system to the local point cloud map in the preset second coordinate system with reference to the position information in the preset first coordinate system includes:

[0029] In the preset first coordinate system, obtain the positional relationship between the position information and the coordinates of each point data;

[0030] Project the position information in the preset first coordinate system to the local point cloud map in the preset second coordinate system as a reference coordinate based on a preset projection rule;

[0031] Project the coordinates of each point data in the preset first coordinate system to the local point cloud map in the preset second coordinate system based on the reference coordinate and the positional relationship.

[0032] Optionally, the preset projection rule includes the following formula:

[0033] P_w = T1 × T2 × p_l;

[0034] where P_w represents the reference coordinate in the preset second coordinate system, p_l represents the position information in the preset first coordinate system, T1 is a transformation matrix obtained through the position information, and T2 is a transformation matrix obtained through the external parameters between the lidar in the intelligent vehicle and the vehicle body coordinate system of the intelligent vehicle.

[0035] According to a specific embodiment of the present disclosure, in a second aspect, the present disclosure provides an intelligent mobile device, including:

[0036] A multi-sensor unit configured to collect the position information and frame point cloud of the intelligent vehicle;

[0037] A processor communicatively connected to the multi-sensor unit and configured to:

[0038] Obtain the position information and frame point cloud of the intelligent vehicle, where the frame point cloud includes a plurality of point data, and both the position information and the plurality of point data are set in a preset first coordinate system;

[0039] Intercept a local point cloud map in the preset point cloud map based on the position information, where the local point cloud map is set in a preset second coordinate system and includes a plurality of grids;

[0040] Based on the local point cloud map, obtain the horizontal projection area where each vertical void space in any grid is projected onto a preset horizontal plane;

[0041] With the position information in the preset first coordinate system as a reference, project the coordinates of each point data in the preset first coordinate system into the local point cloud map in the preset second coordinate system, and obtain the horizontal projection coordinates of each point data projected onto the preset horizontal plane;

[0042] When the horizontal projection coordinates of any point data exist in any horizontal projection area, determine that the point data corresponding to the horizontal projection coordinates is point cloud noise.

[0043] Optionally, the processor is further configured to:

[0044] While obtaining the horizontal projection coordinates of each point data projected onto the preset horizontal plane, also obtain the grid coordinate index of the grid to which each point data belongs in the preset second coordinate system, where the grid coordinate index refers to the coordinates of the grid on the coordinate plane formed by the first coordinate axis and the second coordinate axis in the preset second coordinate system.

[0045] Optionally, the processor is configured to obtain the grid coordinate index of the grid to which each point data belongs in the preset second coordinate system, including:

[0046] Obtain the first coordinate of any point data in the direction of the first coordinate axis, and

[0047] Obtain the second coordinate of the point data in the direction of the second coordinate axis;

[0048] Calculate the quotient of the first coordinate and the preset length of the grid in the direction of the first coordinate axis to obtain the index first coordinate of the grid coordinate index in the direction of the first coordinate axis;

[0049] Calculate the quotient of the second coordinate and the preset width of the grid in the direction of the second coordinate axis to obtain the index second coordinate of the grid coordinate index in the direction of the second coordinate axis.

[0050] Optionally, when the horizontal projection coordinates of any point data exist in any horizontal projection area, the processor is configured to determine that the point data corresponding to the horizontal projection coordinates is point cloud noise, including:

[0051] Obtain all horizontal projection areas of the grid based on the grid coordinate index of the grid to which any point data belongs;

[0052] Traverse all horizontal projection areas of the grid based on the horizontal projection coordinates of the point data;

[0053] When the horizontal projection coordinates of the point data exist in any horizontal projection area of the grid, determine that the point data is point cloud noise.

[0054] Optionally, each horizontal projection area is a rectangular projection, and two adjacent sides of the rectangular projection are respectively parallel to the first coordinate axis and the second coordinate axis in a preset second coordinate system;

[0055] Correspondingly, when the horizontal projection coordinates of the point data exist in any horizontal projection area of the grid, the processor is configured to determine that the point data is point cloud noise, including:

[0056] Obtain the minimum value and the maximum value of any horizontal projection area on the first coordinate axis, generate a first value range of the horizontal projection area on the first coordinate axis, and

[0057] Obtain the minimum value and the maximum value of the horizontal projection area on the second coordinate axis, generate a second value range of the horizontal projection area on the second coordinate axis;

[0058] When the value of the horizontal projection coordinates of the point data on the first coordinate axis is within the first value range and the value of the horizontal projection coordinates of the point data on the second coordinate axis is within the second value range, determine that the point data is point cloud noise.

[0059] Optionally, the processor is configured to project the coordinates of each point data in a preset first coordinate system onto a local point cloud map in a preset second coordinate system with reference to the position information in the preset first coordinate system, including:

[0060] In the preset first coordinate system, obtain the positional relationship between the position information and the coordinates of each point data;

[0061] Project the position information in the preset first coordinate system onto the local point cloud map in the preset second coordinate system as a reference coordinate based on a preset projection rule;

[0062] Project the coordinates of each point data in the preset first coordinate system onto the local point cloud map in the preset second coordinate system based on the reference coordinate and the positional relationship.

[0063] Optionally, the preset projection rule includes the following formula:

[0064] P_w = T1 × T2 × p_l;

[0065] Wherein, P_w represents the reference coordinate in the preset second coordinate system, p_l represents the position information in the preset first coordinate system, T1 is a transformation matrix obtained through the position information, and T2 is a transformation matrix obtained through the external parameters in the lidar of the intelligent vehicle and the body coordinate system of the intelligent vehicle.

[0066] Compared with the prior art, the above solution of the embodiments of the present disclosure has at least the following beneficial effects:

[0067] The present disclosure provides a method for determining point cloud noise and an intelligent mobile device. The present disclosure uses a pre-established preset point cloud map as prior information, and uses the vertical void space in the grid of the preset point cloud map to test the point data in the frame point cloud obtained by real-time scanning, so as to remove the useless points and / or noise point data in the frame point cloud. The present disclosure does not need to set a discrimination threshold, is not affected by environmental changes, and has a stable filtering effect. It avoids the statistics of a large amount of data, can be completed by using grid coordinate indexing and comparison of spatial regions, reduces the data processing volume, and improves the efficiency of identifying noise. At the same time, it does not rely on model training and online deployment, and does not rely on the support of high-performance hardware, reducing the usage cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 Shows a flowchart of a method for determining point cloud noise according to an embodiment of the present disclosure;

[0069] Figure 2 Shows a schematic diagram of a grid in a preset second coordinate system according to an embodiment of the present disclosure;

[0070] Figure 3 Shows a schematic structural diagram of an intelligent mobile device for point cloud noise according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0071] In order to make the objectives, technical solutions, and advantages of the present disclosure clearer, the present disclosure will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present disclosure without creative efforts shall fall within the protection scope of the present disclosure.

[0072] The terms used in the embodiments of the present disclosure are for the purpose of describing specific embodiments only and are not intended to limit the present disclosure. The singular forms "a", "the" and "said" used in the embodiments of the present disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. "Plural" generally includes at least two.

[0073] It should be understood that the term "and / or" used herein is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after.

[0074] It should be understood that although terms such as first, second, and third may be used in the embodiments of the present disclosure for description, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, without departing from the scope of the embodiments of the present disclosure, the first may also be referred to as the second, and similarly, the second may also be referred to as the first.

[0075] Depending on the context, the words "if", "when" as used herein can be interpreted as "when...", "when...", "in response to determining", or "in response to detecting". Similarly, depending on the context, the phrase "if determined" or "if detecting (stated condition or event)" can be interpreted as "when determined", "in response to determining", "when detecting (stated condition or event)", or "in response to detecting (stated condition or event)".

[0076] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a commodity or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such commodity or device. Without further limitation, the element defined by the statement "including one..." does not exclude the existence of additional identical elements in the commodity or device including the said element.

[0077] It should be particularly noted that the symbols and / or numbers existing in the specification, if not marked in the figure description, are not figure labels.

[0078] The optional embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings.

[0079] Embodiment 1

[0080] For the embodiment provided by the present disclosure, that is, an embodiment of a method for determining point cloud noise.

[0081] Point cloud noise includes a large number of floating noise points generated when the laser emitted by the lidar hits rain, snow, fog, etc. under special weather conditions such as rain, snow, and fog. Point cloud noise is laser scan points generated by non-fixed objects around the intelligent vehicle that have no positioning value, or point data in the point cloud that is disadvantageous for positioning brought by the surrounding environment.

[0082] The following will combine Figure 1 to elaborate on the embodiments of the present disclosure in detail.

[0083] Step S101, obtain the position information and frame point cloud of the intelligent vehicle.

[0084] The intelligent vehicle obtains the position information and frame point cloud of the intelligent vehicle through a multi-sensor unit provided inside it.

[0085] The multi-sensor unit collects data collected by a variety of sensors. The multi-sensor unit includes: a global navigation satellite unit, a chassis data acquisition unit, an inertial measurement unit, a camera, and / or a lidar. Among them, the global navigation satellite unit is used to collect position information, the inertial measurement unit is used to collect inertial information, and the lidar is used to collect the frame point cloud. Through an algorithm program, the position information, attitude information, and / or speed value of the intelligent vehicle can be obtained in real time based on the data collected by the multi-sensor unit.

[0086] Point cloud (English full name: point cloud data) refers to a collection of a large amount of point data representing the surface characteristics of an object in a three-dimensional coordinate system, and each point data includes a set of vectors.

[0087] The frame point cloud is the point cloud obtained after the lidar scans the surrounding environment for one week. The frame point cloud includes a plurality of point data, and the point data includes ranging information and intensity information.

[0088] The position information and the plurality of point data are both set in a preset first coordinate system. For example, the preset first coordinate system is the world coordinate system.

[0089] Step S102, intercept a local point cloud map based on the position information in the preset point cloud map.

[0090] The preset point cloud map is a three-dimensional electronic map pre-stitched by calculating the relative poses of multiple frames of point clouds. The preset point cloud map is used as the current basis to assist the safe driving of the intelligent vehicle. Optionally, the preset point cloud map is a high-precision point cloud map.

[0091] The local point cloud map is a partial point cloud map intercepted from the preset point cloud map. The projection coordinates of the position information are included in the local point cloud map. For example, a cube with a preset side length is intercepted from the preset point cloud map with the projection coordinates of the position information as the geometric center as the local point cloud map.

[0092] Wherein, both the preset point cloud map and the local point cloud map are set in a preset second coordinate system and include a plurality of grids. For example, the preset second coordinate system refers to the map coordinate system.

[0093] Step S103: Obtain the horizontal projection area of each vertical void space projected onto a preset horizontal plane in any grid based on the local point cloud map.

[0094] A grid is also called a voxel. The grid processing, also known as voxelization processing, is to divide the three-dimensional space into three-dimensional grids in the preset coordinate system and normalize the point data in the three-dimensional grids, thereby reducing the amount of data operation. After the grid processing, there is point data in some grids and no point data in some grids.

[0095] That is, the preset point cloud map has been subjected to grid processing, and each element information in the preset point cloud map is divided into grids for management. Through grid management, the convergence speed of data processing can be accelerated, and at the same time, the processing of complex and large data is simplified into the processing of grids, improving the efficiency of data processing.

[0096] Such as Figure 2 As shown, the vertical void space forms a column perpendicular to the preset horizontal plane in the grid of the local point cloud map, and the column takes two opposite faces of the grid as the top and bottom surfaces. There is no element information of any local point cloud map in the space included in the column. That is, the projection of the column on the preset horizontal plane does not include the projection coordinates of any element information in the grid. There is no element information in some grids, so the entire grid is a vertical void space, and there is element information in some other grids, so there is at least one vertical void space.

[0097] The preset second coordinate system includes at least a first coordinate axis and a second coordinate axis. The coordinate axis horizontal plane formed by the first coordinate axis and the second coordinate axis. For example, the preset second coordinate system is the XYZ coordinate system, the first coordinate axis is the X axis, the second coordinate axis is the Y axis, and the coordinate axis horizontal plane is the XY plane. The preset horizontal plane can be parallel to the coordinate axis horizontal plane or include the coordinate axis horizontal plane, which is not limited in this disclosure.

[0098] Step S104: With the position information in the preset first coordinate system as a reference, project the coordinates of each point data in the preset first coordinate system onto the local point cloud map in the preset second coordinate system, and obtain the horizontal projection coordinates of each point data projected onto the preset horizontal plane.

[0099] In some specific embodiments, the step of projecting the coordinates of each point data in the preset first coordinate system onto the local point cloud map in the preset second coordinate system with the position information in the preset first coordinate system as a reference includes the following steps:

[0100] Step S104-1: In the preset first coordinate system, obtain the positional relationship between the position information and the coordinates of each point data.

[0101] The positional relationship includes the first relative difference between the position information and the coordinates of the point data on the first coordinate axis and the second relative difference between the position information and the coordinates of the point data on the second coordinate axis. The first relative difference or the second relative difference can be either negative or positive. For example, on the first coordinate axis, if the coordinate of the position information is 8 and the coordinate of the first point data is 6, then the first relative difference is 6 - 8 = -2; if the coordinate of the position information is 8 and the coordinate of the second point data is 15, then the first relative difference is 15 - 8 = 7.

[0102] Step S104-2: Project the position information in the preset first coordinate system onto the local point cloud map in the preset second coordinate system as a reference coordinate based on the preset projection rule.

[0103] Specifically, the preset projection rule includes the following formula:

[0104] P_w = T1 × T2 × p_l;

[0105] where P_w represents the reference coordinate in the preset second coordinate system, p_l represents the position information in the preset first coordinate system, T1 is a transformation matrix obtained through the position information, and T2 is a transformation matrix obtained through the external parameters between the lidar in the intelligent vehicle and the vehicle body coordinate system of the intelligent vehicle.

[0106] Step S104-3: Project the coordinates of each point data in the preset first coordinate system onto the local point cloud map in the preset second coordinate system based on the reference coordinate and the positional relationship.

[0107] After the position information in the preset first coordinate system determines its position in the local point cloud map in the preset second coordinate system, through

[0108] In some other specific embodiments, the method further includes the following steps:

[0109] Step S104a, while obtaining the horizontal projection coordinates of each point data projected onto a preset horizontal plane, also obtain the grid coordinate index of the grid to which each point data belongs in a preset second coordinate system.

[0110] Wherein, the grid coordinate index refers to the coordinates of the grid on the coordinate axis horizontal plane formed by the first coordinate axis and the second coordinate axis in the preset second coordinate system.

[0111] Specifically, the obtaining of the grid coordinate index of the grid to which each point data belongs in the preset second coordinate system includes the following steps:

[0112] Step S104a-1, obtain the first coordinate of any point data in the direction of the first coordinate axis, and obtain the second coordinate of the point data in the direction of the second coordinate axis.

[0113] For example, the first coordinate of the first point data in the direction of the first coordinate axis is 20, and the second coordinate of the point data in the direction of the second coordinate axis is 18, that is, the coordinates of the first point data are (20, 18).

[0114] Step S104a-2, calculate the quotient of the first coordinate and the preset length of the grid in the direction of the first coordinate axis to obtain the index first coordinate of the grid coordinate index in the direction of the first coordinate axis.

[0115] For example, continuing the above example, if the coordinates of the first point data are in the first grid, and the preset length of the first grid in the direction of the first coordinate axis is 6, then the index first coordinate of the first grid is 20÷6 = 3.

[0116] Step S104a-3, calculate the quotient of the second coordinate and the preset width of the grid in the direction of the second coordinate axis to obtain the index second coordinate of the grid coordinate index in the direction of the second coordinate axis.

[0117] For example, continuing the above example, if the preset width of the first grid in the direction of the second coordinate axis is 4, then the index second coordinate of the first grid is 18÷4 = 4. Therefore, the grid coordinate index of the first grid is (3, 4).

[0118] Step S105, when the horizontal projection coordinates of any point data exist in any horizontal projection area, determine that the point data corresponding to the horizontal projection coordinates is point cloud noise.

[0119] In some specific embodiments, the determining that the point data corresponding to the horizontal projection coordinates is point cloud noise when the horizontal projection coordinates of any point data exist in any horizontal projection area includes the following steps:

[0120] Step S105-1: Obtain all horizontal projection regions of the grid based on the grid coordinate index of the grid to which any point data belongs.

[0121] For example, continuing with the above example, the grid coordinate index (3, 4) of the first grid where the first point data is located is obtained through the coordinates of the first point data. The first grid can be retrieved through the grid coordinate index (3, 4) of the first grid, and thus all horizontal projection regions in the first grid can be obtained.

[0122] Step S105-2: Traverse all horizontal projection regions of the grid based on the horizontal projection coordinates of the point data.

[0123] Step S105-3: When the horizontal projection coordinates of the point data exist in any horizontal projection region of the grid, determine that the point data is point cloud noise.

[0124] In the embodiments of the present disclosure, a preset point cloud map is used as prior information to determine whether point data is point cloud noise. Since the horizontal projection region is the horizontal projection region of the vertical void space on the preset horizontal plane in the local point cloud map, that is, there is no point data in the vertical void space in the local point cloud map. If the point data in the frame point cloud appears in this vertical void space, which is manifested as the horizontal projection coordinates of the point data existing in any horizontal projection region of the grid, it indicates that the point data is point cloud noise.

[0125] In some specific embodiments, as Figure 2 shown, each horizontal projection region is a rectangular projection, and two adjacent sides of the rectangular projection are respectively parallel to the first coordinate axis and the second coordinate axis in the preset second coordinate system.

[0126] Correspondingly, when the horizontal projection coordinates of the point data exist in any horizontal projection region of the grid, determining that the point data is point cloud noise includes the following steps:

[0127] Step S105-3-1: Obtain the minimum value and the maximum value of any horizontal projection region on the first coordinate axis, generate the first value range of the horizontal projection region on the first coordinate axis, and obtain the minimum value and the maximum value of the horizontal projection region on the second coordinate axis, generate the second value range of the horizontal projection region on the second coordinate axis.

[0128] For example, as Figure 2 shown, if the minimum value and the maximum value of the horizontal projection region on the first coordinate axis are 4 and 6 respectively, the first value range is [4, 6]; if the minimum value and the maximum value of the horizontal projection region on the second coordinate axis are 7 and 10 respectively, the second value range is [7, 10].

[0129] Step S105-3-2, when the value of the horizontal projection coordinate of the point data on the first coordinate axis is within the first value range, and the value of the horizontal projection coordinate of the point data on the second coordinate axis is within the second value range, the point data is determined to be a point cloud noise point.

[0130] For example, continuing with the above example, if the horizontal projection coordinates of the point data are (4.8, 9.2), 4.8∈the first value range [4, 6], 9.2∈the second value range [7, 10], then the point data is determined to be a point cloud noise point.

[0131] The disclosed embodiment uses a pre-established preset point cloud map as prior information, and utilizes the vertical hole space in the grid of the preset point cloud map to check the point data in the frame point cloud obtained by real-time scanning, so as to remove useless points and / or noise point data in the frame point cloud. The disclosed embodiment does not need to set a discrimination threshold, is not affected by environmental changes, and has a stable filtering effect. It avoids the statistics of massive data and can be completed by comparing grid coordinate indexes and spatial areas, reducing the amount of data processing and improving the efficiency of identifying noise points. At the same time, it does not rely on model training and online deployment, and does not rely on the support of high-performance hardware, which reduces the cost of use.

[0132] Example 2

[0133] The present disclosure also provides an apparatus embodiment that is consistent with the above-mentioned embodiment, and is used to implement the method steps described in the above-mentioned embodiment. The explanation based on the same name meaning is the same as that of the above-mentioned embodiment, and has the same technical effect as that of the above-mentioned embodiment, and will not be repeated here.

[0134] like Figure 3 As shown, the present disclosure provides a smart mobile device 300, including:

[0135] The multi-sensor unit 301 is configured to collect position information and frame point cloud of the intelligent vehicle;

[0136] The processor 302 is in communication with the multi-sensor unit 301 and is configured to:

[0137] Acquire position information and a frame point cloud of the intelligent vehicle, wherein the frame point cloud includes a plurality of point data, and the position information and the plurality of point data are both set in a preset first coordinate system;

[0138] intercepting a local point cloud map based on the location information in the preset point cloud map, wherein the local point cloud map is set in a preset second coordinate system and includes a plurality of grids;

[0139] Based on the local point cloud map, obtain the horizontal projection areas of each vertical hole space in any grid projected onto a preset horizontal plane;

[0140] Taking the position information in the preset first coordinate system as a reference, project the coordinates of each point data in the preset first coordinate system onto the local point cloud map in the preset second coordinate system, and obtain the horizontal projection coordinates of each point data projected onto the preset horizontal plane;

[0141] When the horizontal projection coordinates of any point data exist in any horizontal projection area, determine that the point data corresponding to the horizontal projection coordinates is point cloud noise.

[0142] Optionally, the processor 302 is further configured to:

[0143] While obtaining the horizontal projection coordinates of each point data projected onto the preset horizontal plane, also obtain the grid coordinate index of the grid to which each point data belongs in the preset second coordinate system, where the grid coordinate index refers to the coordinates of the grid on the coordinate plane formed by the first coordinate axis and the second coordinate axis in the preset second coordinate system.

[0144] Optionally, the processor 302 is configured to obtain the grid coordinate index of the grid to which each point data belongs in the preset second coordinate system, including:

[0145] Obtain the first coordinate of any point data in the direction of the first coordinate axis, and

[0146] Obtain the second coordinate of the point data in the direction of the second coordinate axis;

[0147] Calculate the quotient of the first coordinate and the preset length of the grid in the direction of the first coordinate axis to obtain the index first coordinate of the grid coordinate index in the direction of the first coordinate axis;

[0148] Calculate the quotient of the second coordinate and the preset width of the grid in the direction of the second coordinate axis to obtain the index second coordinate of the grid coordinate index in the direction of the second coordinate axis.

[0149] Optionally, the processor 302 is configured to, when the horizontal projection coordinates of any point data exist in any horizontal projection area, determine that the point data corresponding to the horizontal projection coordinates is point cloud noise, including:

[0150] Obtain all horizontal projection areas of the grid based on the grid coordinate index of the grid to which the point data belongs;

[0151] Traverse all horizontal projection areas of the grid based on the horizontal projection coordinates of the point data;

[0152] When the horizontal projection coordinates of the point data exist in any horizontal projection area of the grid, determine that the point data is point cloud noise.

[0153] Optionally, each horizontal projection area is a rectangular projection, and two adjacent sides of the rectangular projection are respectively parallel to the first coordinate axis and the second coordinate axis in a preset second coordinate system;

[0154] Correspondingly, the processor 302 is configured to determine that the point data is point cloud noise when the horizontal projection coordinates of the point data exist in any horizontal projection area of the grid, including:

[0155] Obtain the minimum value and the maximum value of any horizontal projection area on the first coordinate axis, generate a first value range of the horizontal projection area on the first coordinate axis, and

[0156] Obtain the minimum value and the maximum value of the horizontal projection area on the second coordinate axis, generate a second value range of the horizontal projection area on the second coordinate axis;

[0157] When the value of the horizontal projection coordinates of the point data on the first coordinate axis is within the first value range, and the value of the horizontal projection coordinates of the point data on the second coordinate axis is within the second value range, determine that the point data is point cloud noise.

[0158] Optionally, the processor 302 is configured to project the coordinates of each point data in the preset first coordinate system to the local point cloud map in the preset second coordinate system with reference to the position information in the preset first coordinate system, including:

[0159] In the preset first coordinate system, obtain the positional relationship between the position information and the coordinates of each point data;

[0160] Project the position information in the preset first coordinate system to the local point cloud map in the preset second coordinate system as a reference coordinate based on a preset projection rule;

[0161] Project the coordinates of each point data in the preset first coordinate system to the local point cloud map in the preset second coordinate system based on the reference coordinate and the positional relationship.

[0162] Optionally, the preset projection rule includes the following formula:

[0163] P_w = T1 × T2 × p_l;

[0164] where P_w represents the reference coordinate in the preset second coordinate system, p_l represents the position information in the preset first coordinate system, T1 is a transformation matrix obtained through the position information, and T2 is a transformation matrix obtained through the external parameters between the lidar in the intelligent vehicle and the vehicle body coordinate system of the intelligent vehicle.

[0165] Embodiments of the present disclosure use a pre-established preset point cloud map as prior information, and use the vertical void space in the grid of the preset point cloud map to check the point data in the frame point cloud obtained by real-time scanning, so as to remove the useless points and / or the point data of noise points in the frame point cloud. Embodiments of the present disclosure do not need to set a discrimination threshold, are not affected by environmental changes, and have a stable filtering effect. Avoiding the statistics of massive data, it can be completed by using grid coordinate indexing and comparison of spatial regions, reducing the amount of data processing and improving the efficiency of identifying noise points. At the same time, it does not rely on model training and online deployment, and does not rely on the support of high-performance hardware, reducing the usage cost.

[0166] Finally, it should be noted that the embodiments in this specification are described in a progressive manner, and the key points of each embodiment are the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the systems or devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and the relevant parts can be referred to the description of the method part.

[0167] The above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them; although the present disclosure has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure.

Claims

1. A method for determining noise points in a point cloud, characterized in that Including: Obtain the position information and frame point cloud of the intelligent vehicle, where the frame point cloud includes a plurality of point data, and both the position information and the plurality of point data are set in a preset first coordinate system; Intercept a local point cloud map based on the position information in the preset point cloud map, where the local point cloud map is set in a preset second coordinate system and includes a plurality of grids; Obtain the horizontal projection area of each vertical void space in any grid projected onto a preset horizontal plane based on the local point cloud map; Project the coordinates of each point data in the preset first coordinate system to the local point cloud map in the preset second coordinate system with reference to the position information in the preset first coordinate system, and obtain the horizontal projection coordinates of each point data projected onto the preset horizontal plane; When the horizontal projection coordinates of any point data exist in any horizontal projection area, determine that the point data corresponding to the horizontal projection coordinates is point cloud noise.

2. The method according to claim 1, wherein The method further includes: While obtaining the horizontal projection coordinates of each point data projected onto the preset horizontal plane, also obtain the grid coordinate index of the grid to which each point data belongs in the preset second coordinate system, where the grid coordinate index refers to the coordinates of the grid on the coordinate plane formed by the first coordinate axis and the second coordinate axis in the preset second coordinate system.

3. The method according to claim 2, wherein The obtaining the grid coordinate index of the grid to which each point data belongs in the preset second coordinate system includes: Obtain the first coordinate of any point data in the direction of the first coordinate axis, and Obtain the second coordinate of the point data in the direction of the second coordinate axis; Calculate the quotient of the first coordinate and the preset length of the grid in the direction of the first coordinate axis to obtain the index first coordinate of the grid coordinate index in the direction of the first coordinate axis; Calculate the quotient of the second coordinate and the preset width of the grid in the direction of the second coordinate axis to obtain the index second coordinate of the grid coordinate index in the direction of the second coordinate axis.

4. The method according to claim 2, characterized in that The when the horizontal projection coordinates of any point data exist in any horizontal projection area, determine that the point data corresponding to the horizontal projection coordinates is point cloud noise, includes: Obtain all the horizontal projection areas of the grid based on the grid coordinate index of the grid to which any point data belongs; Traverse all the horizontal projection areas of the grid based on the horizontal projection coordinates of the point data; When the horizontal projection coordinates of the point data exist in any horizontal projection area of the grid, determine that the point data is point cloud noise.

5. The method according to claim 4, wherein Each horizontal projection area is a rectangular projection, and two adjacent sides of the rectangular projection are respectively parallel to the first coordinate axis and the second coordinate axis in the preset second coordinate system; Correspondingly, the when the horizontal projection coordinates of the point data exist in any horizontal projection area of the grid, determine that the point data is point cloud noise, includes: Obtain the minimum value and the maximum value of any horizontal projection area on the first coordinate axis, generate the first value range of the horizontal projection area on the first coordinate axis, and Obtain the minimum and maximum values of the horizontal projection area on the second coordinate axis, and generate a second value range of the horizontal projection area on the second coordinate axis; When the value of the horizontal projection coordinate of the point data on the first coordinate axis is within the first value range, and the value of the horizontal projection coordinate of the point data on the second coordinate axis is within the second value range, determine that the point data is point cloud noise.

6. The method according to claim 1, characterized in that, The projection of the coordinates of each point data in the preset first coordinate system onto the local point cloud map in the preset second coordinate system with reference to the position information in the preset first coordinate system includes: In the preset first coordinate system, obtain the positional relationship between the position information and the coordinates of each point data; Project the position information in the preset first coordinate system onto the local point cloud map in the preset second coordinate system based on a preset projection rule as a reference coordinate; Based on the reference coordinate and the positional relationship, project the coordinates of each point data in the preset first coordinate system onto the local point cloud map in the preset second coordinate system.

7. The method according to claim 6, characterized in that, The preset projection rule includes the following formula: P_w = T1 × T2 × p_l; where P_w represents the reference coordinate in the preset second coordinate system, p_l represents the position information in the preset first coordinate system, T1 is a transformation matrix obtained through the position information, and T2 is a transformation matrix obtained through the external parameters between the lidar in the intelligent vehicle and the vehicle body coordinate system of the intelligent vehicle.

8. An intelligent mobile device, characterized in that, including: A multi-sensor unit configured to collect the position information and frame point cloud of the intelligent vehicle; A processor communicatively connected to the multi-sensor unit and configured to: Obtain the position information and frame point cloud of the intelligent vehicle, where the frame point cloud includes a plurality of point data, and both the position information and the plurality of point data are set in a preset first coordinate system; Intercept a local point cloud map in the preset point cloud map based on the position information, where the local point cloud map is set in a preset second coordinate system and includes a plurality of grids; Based on the local point cloud map, obtain the horizontal projection area of each vertical void space projected onto a preset horizontal plane in any grid; With reference to the position information in the preset first coordinate system, project the coordinates of each point data in the preset first coordinate system onto the local point cloud map in the preset second coordinate system, and obtain the horizontal projection coordinates of each point data projected onto the preset horizontal plane; When the horizontal projection coordinate of any point data exists in any horizontal projection area, determine that the point data corresponding to the horizontal projection coordinate is point cloud noise.

9. The device according to claim 8, characterized in that, The processor is further configured to: When obtaining the horizontal projection coordinates of each point data projected onto the preset horizontal plane, also obtain the grid coordinate index of the grid to which each point data belongs in the preset second coordinate system, where the grid coordinate index refers to the coordinates of the grid on the coordinate plane formed by the first coordinate axis and the second coordinate axis in the preset second coordinate system.

10. The device according to claim 9, characterized in that, The processor is configured to obtain the grid coordinate index of the grid to which each point data belongs in the preset second coordinate system, including: Obtain the first coordinate of any point data in the direction of the first coordinate axis, and Obtain the second coordinate of the point data in the direction of the second coordinate axis; Calculate the quotient of the first coordinate and the preset length of the grid in the direction of the first coordinate axis to obtain the index first coordinate of the grid coordinate index in the direction of the first coordinate axis; Calculate the quotient of the second coordinate and the preset width of the grid in the direction of the second coordinate axis to obtain the index second coordinate of the grid coordinate index in the direction of the second coordinate axis.

Images