Laser radar point cloud abnormal point removal method and device, electronic equipment and storage medium

Abstract

The invention provides a laser radar point cloud abnormal point removal method and device, electronic equipment and a storage medium, and relates to the technical field of laser radar data processing, and the method comprises the steps: firstly obtaining the point cloud data of each frame of a laser radar; and then traversing each frame of point cloud data, determining an obstacle distance in the current frame of point cloud data, determining candidate abnormal points of an obstacle edge in the current frame of point cloud data based on the obstacle distance, determining a target abnormal point based on the candidate abnormal points, and removing the target abnormal point. According to the method, the number of candidate abnormal points is dynamically adjusted according to the obstacle distance, different abnormal points of different distances can be covered, the target abnormal point is removed from each frame of point cloud data, the accuracy and reliability of each frame of point cloud data are improved, and the influence of the target abnormal point on each frame of point cloud data is effectively reduced; and the accuracy and precision of the downstream task result of the point cloud data are improved. When the laser radar is applied to the robot, the collision risk of the robot can be reduced, and the obstacle avoidance capability of the robot is improved.

Description

Technical Field

[0001] The present invention relates to the field of laser radar data processing technology, and in particular to a laser radar point cloud outlier removal method, device, electronic device and storage medium. Background Art

[0002] With the rapid advancement of robotics technology, Simultaneous Localization and Mapping (SLAM) technology has been widely used in various industries, especially in the use of lidar for environmental perception, mapping and positioning, and autonomous navigation, which has shown great potential. This has put forward more stringent requirements on the quality of point cloud data collected by lidar.

[0003] In practical applications, a significant challenge facing LiDAR is that when encountering close obstacles, the reflection behavior of the laser pulse becomes complicated: if part of the laser pulse happens to hit the edge of the obstacle, and the remaining laser pulse continues to project to the rear wall, the two parts of the reflected pulse signals may be superimposed on each other, resulting in deviations in the laser ranging results, erroneously shortening the distance between the obstacle and the wall, and causing distortion of the point cloud data. This distortion not only affects the accuracy of map construction, but also reduces the accuracy of the positioning system, posing a challenge to the robot's autonomous navigation capabilities. Moreover, when the distance between the obstacle and the LiDAR is extremely close, the radar system's own noise interference will also increase, making the point cloud data in the area around the obstacle even more inaccurate. These inaccurate data points may mislead the robot's path planning algorithm, increase the risk of collision or cause navigation failure.

[0004] In order to make the point cloud data obtained by the laser radar more accurate, the current technology mainly uses distance thresholds and angle thresholds to judge and remove the trailing points in the point cloud data. The process first calculates the distance between each point and its left and right adjacent points. If both distance values ​​exceed the distance threshold, the point is initially regarded as a potential trailing point. Next, for the initially determined trailing points, they are further verified by calculating the inner angles of two specific triangles: one is the angle corresponding to the point cloud origin in the triangle formed by the point, its left adjacent point and the point cloud origin, and the other is the angle corresponding to the point cloud origin in the triangle formed by the point, its right adjacent point and the point cloud origin. If both angles are less than the angle threshold, the point is finally confirmed as a trailing point and filtered out. Finally, the point clouds generated by each laser beam after the trailing point filtering process are superimposed to synthesize a frame of complete point cloud data that does not contain trailing points. However, this method can only be used to remove trailing points in point cloud data, and cannot remove other abnormal points in point cloud data, and still cannot guarantee the accuracy of point cloud data. Summary of the invention

[0005] The present invention provides a method, device, electronic device and storage medium for removing abnormal points of a laser radar point cloud, so as to solve the defects existing in the related art.

[0006] The present invention provides a method for removing abnormal points in a laser radar point cloud, comprising: Get each frame of point cloud data of the laser radar; Traversing each frame of point cloud data, determining the obstacle distance in the current frame of point cloud data, determining candidate outlier points at the edge of the obstacle in the current frame of point cloud data based on the obstacle distance, determining target outlier points based on the candidate outlier points, and removing the target outlier points.

[0007] According to a method for removing abnormal points from a laser radar point cloud provided by the present invention, the method of determining candidate abnormal points at the edge of an obstacle in the point cloud data of the current frame based on the obstacle distance comprises: Extracting obstacle edge points in the current frame point cloud data; Based on the obstacle distance, determining an edge outlier point adjacent to the obstacle edge point; The candidate outlier point is determined based on the obstacle edge point and the edge outlier point.

[0008] According to a method for removing abnormal points from a laser radar point cloud provided by the present invention, the method of determining an edge abnormal point adjacent to the obstacle edge point based on the obstacle distance comprises: Determine a target distance range within which the obstacle distance lies; Based on the correspondence between the distance range and the number of edge outliers, determine the number of target edge outliers corresponding to the target distance range; Determining the edge outliers based on the number of the target edge outliers; Among them, in the corresponding relationship, a distance range with a large value corresponds to a small number of edge outliers.

[0009] According to a method for removing abnormal points from a laser radar point cloud provided by the present invention, the corresponding relationship includes a first distance range, a second distance range, and a third distance range whose values ​​increase in sequence; The number of edge outliers corresponding to the first distance range is greater than the number of edge outliers corresponding to the third distance range, and each value within the second distance range is inversely proportional to the corresponding number of edge outliers.

[0010] According to a method for removing abnormal points from a laser radar point cloud provided by the present invention, the method of determining a target abnormal point based on the candidate abnormal point comprises: Performing straight line fitting on the point cloud data of the current frame to obtain a plurality of fitting straight lines; Based on the vertical distances from the candidate outlier points to each fitting straight line, the candidate outlier points are screened to obtain the target outlier points.

[0011] According to a method for removing abnormal points from a laser radar point cloud provided by the present invention, the candidate abnormal points are screened based on the vertical distances from the candidate abnormal points to each fitting straight line to obtain the target abnormal point, including: For any candidate outlier point, determine the minimum vertical distance from any candidate outlier point to each of the fitted straight lines; If the minimum vertical distance is greater than or equal to a preset threshold, any of the candidate abnormal points is determined to be the target abnormal point.

[0012] The present invention also provides a laser radar point cloud abnormal point removal device, comprising: A data acquisition module is used to obtain point cloud data of each frame of the laser radar; The outlier removal module is used to traverse the point cloud data of each frame, determine the obstacle distance in the current frame point cloud data, determine the candidate outlier points of the obstacle edge in the current frame point cloud data based on the obstacle distance, determine the target outlier points based on the candidate outlier points, and remove the target outlier points.

[0013] The present invention also provides an electronic device, including a laser radar, a memory, a processor, and a computer program stored in the memory and runnable on the processor, wherein the laser radar is used to collect point cloud data of each frame, and when the processor executes the program, a laser radar point cloud outlier removal method as described in any one of the above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for removing abnormal points from a laser radar point cloud as described in any one of the above is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-described methods for removing abnormal points from a laser radar point cloud.

[0016] The laser radar point cloud outlier removal method, device, electronic device and storage medium provided by the present invention first obtain each frame of laser radar point cloud data; then traverse each frame of point cloud data, determine the obstacle distance in the current frame of point cloud data, determine the candidate outlier points at the edge of the obstacle in the current frame of point cloud data based on the obstacle distance, determine the target outlier points based on the candidate outlier points, and remove the target outlier points. The method can dynamically adjust the number of candidate outlier points according to the obstacle distance, and then cover different outliers at different distances, remove the target outlier points from each frame of point cloud data, improve the accuracy and reliability of each frame of point cloud data, effectively reduce the impact of the target outlier points on each frame of point cloud data, and then improve the accuracy and precision of the downstream task results of the point cloud data. When laser radar is applied to a robot, the robot's collision risk can be reduced and the robot's obstacle avoidance ability can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 This is one of the flow charts of the method for removing abnormal points in a laser radar point cloud provided by the present invention.

[0019] Figure 2 This is the second flow chart of the method for removing abnormal points from a laser radar point cloud provided by the present invention.

[0020] Figure 3 It is a schematic diagram of the current frame point cloud data obtained by laser radar scanning in the laser radar point cloud outlier removal method provided by the present invention.

[0021] Figure 4 It is a schematic diagram of target outlier points determined in the laser radar point cloud outlier point removal method provided by the present invention.

[0022] Figure 5 It is a schematic diagram of the current frame point cloud data obtained after removing abnormal points in the laser radar point cloud removal method provided by the present invention.

[0023] Figure 6 It is a structural schematic diagram of the laser radar point cloud abnormal point removal device provided by the present invention.

[0024] Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] Since the point cloud processing method in the prior art can only remove the tailing points in the point cloud data, but cannot remove other abnormal points in the point cloud data, the accuracy of the point cloud data cannot be guaranteed. Therefore, an embodiment of the present invention provides a method for removing abnormal points in a laser radar point cloud.

[0027] Figure 1 FIG. 1 is a flow chart of a method for removing abnormal points in a laser radar point cloud provided in an embodiment of the present invention. Figure 1 As shown, the method includes: S1, obtain each frame of point cloud data of the laser radar; S2, traversing each frame of point cloud data, determining the obstacle distance in the current frame of point cloud data, determining candidate outlier points at the edge of the obstacle in the current frame of point cloud data based on the obstacle distance, determining target outlier points based on the candidate outlier points, and removing the target outlier points.

[0028] Specifically, the laser radar point cloud outlier removal method provided in the embodiment of the present invention is executed by a laser radar point cloud outlier removal device, which can be configured in a computer, which can be a local computer or a cloud computer. The local computer can be a computer, a tablet, etc., which is not specifically limited here.

[0029] First, step S1 is performed to obtain each frame of point cloud data of the laser radar. The laser radar can collect each frame of point cloud data in time sequence, and each frame of point cloud data can include distance information of each laser radar point, that is, the distance between each point and the laser radar.

[0030] Each laser radar point may have a corresponding number, which may be a non-zero integer. If there are 500 laser radar points in total, each laser radar point may be numbered from 0 to 499 to avoid laser radar point numbering out of bounds.

[0031] The laser radar can be a single-line laser radar, a multi-line laser radar, or other types of laser radar, which is not specifically limited here.

[0032] Then, step S2 is executed to traverse each frame of point cloud data, determine the target abnormal points in each frame of point cloud data, and remove them.

[0033] For the traversed current frame point cloud data, the obstacle distance in the current frame point cloud data can be determined, and the obstacle distance can be the minimum distance in the current frame point cloud data. By using the obstacle distance, candidate abnormal points at the edge of the obstacle in the current frame point cloud data can be screened out.

[0034] It is understandable that the candidate abnormal point may be a neighboring point located at the edge of the obstacle and susceptible to the laser pulse superposition effect, may be a plurality of abnormal points including the obstacle edge point, may include a trailing point, and may include abnormal points near the trailing point.

[0035] The number of candidate outlier points varies with the distance to the obstacle. In the embodiment of the present invention, the greater the distance to the obstacle, the fewer the number of candidate outlier points. The obstacle distance and the number of candidate outlier points may have a negatively correlated functional relationship or a corresponding relationship of specific values, which is determined by actual business needs and is not specifically limited here.

[0036] The target outlier points can be determined by using the candidate outliers. For example, the candidate outliers can be directly removed from the current frame point cloud data as the target outliers, or the candidate outliers can be screened, the normal points among the candidate outliers are removed, and the remaining candidate outliers are removed from the current frame point cloud data as the target outliers, which is not specifically limited here.

[0037] After the traversal is completed, the target outliers can be removed from each frame of point cloud data collected by the laser radar to obtain point cloud data with the target outliers removed from each frame.

[0038] The laser radar point cloud outlier removal method provided in the embodiment of the present invention first obtains each frame of the laser radar point cloud data; then traverses each frame of the point cloud data, determines the obstacle distance in the current frame of the point cloud data, determines the candidate outlier points at the edge of the obstacle in the current frame of the point cloud data based on the obstacle distance, determines the target outlier points based on the candidate outlier points, and removes the target outlier points. The method can dynamically adjust the number of candidate outlier points according to the obstacle distance, and can cover different outliers at different distances. Removing the target outlier points from each frame of the point cloud data can improve the accuracy and reliability of each frame of the point cloud data, effectively reduce the impact of the target outlier points on each frame of the point cloud data, and thus improve the accuracy and precision of the downstream task results of the point cloud data. When the laser radar is applied to the robot, the collision risk of the robot can be reduced and the obstacle avoidance ability of the robot can be improved.

[0039] On the basis of the above embodiment, the step of determining the candidate abnormal point at the edge of the obstacle in the point cloud data of the current frame based on the obstacle distance includes: Extracting obstacle edge points in the current frame point cloud data; Based on the obstacle distance, determining an edge outlier point adjacent to the obstacle edge point; The candidate outlier point is determined based on the obstacle edge point and the edge outlier point.

[0040] Specifically, when determining candidate abnormal points of the obstacle edge in the current frame point cloud data, the obstacle edge points in the current frame point cloud data can be extracted first. When extracting the obstacle edge points, point cloud segmentation can be performed first, for example, first based on height information or plane fitting segmentation, ground points in the current frame point cloud data are removed, and then similar points in the point cloud data from which the ground points are removed are clustered together through clustering algorithms such as density-based spatial clustering of applications with noise (DBSCAN) and K-means to form different clusters, each cluster representing an obstacle.

[0041] After that, edge points in the cluster are extracted as obstacle edge points by edge detection algorithm or geometric feature-based method. Edge detection algorithm can include Canny edge detection, Sobel edge detection, etc., and points with large gradient changes in point cloud data can be extracted as obstacle edge points. Based on geometric features, the contour line of the obstacle can be calculated, and then points on the contour line can be extracted as obstacle edge points. Alternatively, the curvature of the obstacle can be calculated, and points with large curvature can be used as obstacle edge points.

[0042] Thereafter, using the obstacle distance, the edge outlier points adjacent to the obstacle edge point can be determined. The number of edge outliers varies with the obstacle distance. In the embodiment of the present invention, the greater the obstacle distance, the fewer the number of edge outliers. The obstacle distance and the number of edge outliers may have a negatively correlated functional relationship or a corresponding relationship of specific values, which is determined by actual business needs and is not specifically limited here.

[0043] Finally, the obstacle edge points and edge outlier points are used to determine the candidate outlier points. Here, the obstacle edge points and edge outlier points can be directly used as candidate outlier points.

[0044] In an embodiment of the present invention, by determining candidate outlier points by combining obstacle edge points with obstacle distances, the candidate outlier points can include outlier points at the obstacle edge, and then the outlier points at the obstacle edge in the current frame point cloud data can be filtered out, thereby improving the accuracy of the current frame point cloud data.

[0045] Based on the above embodiment, the step of determining the edge outlier point adjacent to the obstacle edge point based on the obstacle distance includes: Determine a target distance range within which the obstacle distance lies; Based on the correspondence between the distance range and the number of edge outliers, determine the number of target edge outliers corresponding to the target distance range; Determining the edge outliers based on the number of the target edge outliers; Among them, in the corresponding relationship, a distance range with a large value corresponds to a small number of edge outliers.

[0046] Specifically, when determining the edge abnormal point adjacent to the obstacle edge point, the target distance range of the obstacle distance can be determined first. Here, multiple continuous distance ranges can be pre-configured, and then by comparing the obstacle distance with the upper and lower limits of each distance range, it is determined which distance range the obstacle distance is in, and the distance range in which the obstacle distance is located is the target distance range.

[0047] It is understandable that each pre-configured distance range may correspond to a number of edge outliers, and each distance range and the corresponding number of edge outliers form a correspondence between the distance range and the number of edge outliers. In this correspondence, a large distance range may correspond to a small number of edge outliers, so that more target outliers may be eliminated for close obstacles, and fewer target outliers may be eliminated for long-distance obstacles, further improving the accuracy of the point cloud data of close obstacles.

[0048] Using this correspondence, the number of target edge outliers corresponding to the target distance range can be determined. Furthermore, the number of target edge outliers can be used to determine the edge outliers. For example, starting from the nearest neighbor points on both sides of the obstacle edge point, the neighboring points of the number of target edge outliers can be selected as edge outliers. If the number of target edge outliers is k, k neighboring points on each side of the obstacle edge point can be selected, and a total of 2k neighboring points on both sides can be selected as edge outliers.

[0049] Afterwards, the obstacle edge points can be combined, and a total of 2k+1 points can be used as candidate outlier points.

[0050] In the embodiment of the present invention, the correspondence between the distance range and the number of edge outliers is combined to determine the number of target edge outliers corresponding to the target distance range, and then determine the edge outliers, which can improve the efficiency of determining the edge outliers.

[0051] On the basis of the above embodiment, the corresponding relationship includes a first distance range, a second distance range and a third distance range whose values ​​increase in sequence; The number of edge outliers corresponding to the first distance range is greater than the number of edge outliers corresponding to the third distance range, and each value within the second distance range is inversely proportional to the corresponding number of edge outliers.

[0052] Specifically, in the correspondence between the distance range and the number of edge outliers, a first distance range, a second distance range, and a third distance range may be included, with values ​​increasing in sequence. For example, the first range may be (0, 0.1), the second distance range may be [0.1, 0.5], and the third distance range may be (0.5, ∞).

[0053] The number of edge outliers corresponding to the first distance range is greater than the number of edge outliers corresponding to the third distance range, and each value in the second distance range is inversely proportional to the corresponding number of edge outliers. For example, the number of edge outliers corresponding to the first distance range may be 10, the number of edge outliers corresponding to the third distance range may be 1, and the number of edge outliers corresponding to each value in the second distance range may be the reciprocal of each value.

[0054] In the embodiment of the present invention, considering that the outliers have little impact on the accuracy and reliability of the point cloud data of distant obstacles or background points, the number of edge outliers corresponding to their distance range can be reduced, thereby reducing the amount of data processing and improving processing efficiency.

[0055] On the basis of the above embodiment, the step of determining a target outlier point based on the candidate outlier point includes: Performing straight line fitting on the point cloud data of the current frame to obtain a plurality of fitting straight lines; Based on the vertical distances from the candidate outlier points to each fitting straight line, the candidate outlier points are screened to obtain the target outlier points.

[0056] Specifically, in an embodiment of the present invention, when using candidate outliers to determine target outliers, a straight line fitting method such as a random sample consensus (RANSAC) algorithm, a Hough transform algorithm, a hierarchical clustering algorithm, a Split-Merge algorithm, etc. can be used to first perform straight line fitting on the current frame point cloud data to extract the edge features of the obstacle and obtain multiple fitting straight lines in the current frame point cloud data.

[0057] There may be multiple candidate outlier points. For each candidate outlier point, its vertical distance to each fitting line can be calculated. Then, based on the vertical distance, it can be determined whether the candidate outlier point can be considered as a normal point within the error range. If it can be considered as a normal point, the candidate outlier point can be eliminated.

[0058] After all candidate outliers are screened, the target outliers are obtained.

[0059] In an embodiment of the present invention, by performing fitting to screen the candidate abnormal points, normal points in the current point cloud data can be retained, and erroneous deletion of valid normal points can be avoided. This can further improve the accuracy and reliability of the current point cloud data and provide higher quality data support for downstream applications.

[0060] On the basis of the above embodiment, the candidate outlier points are screened based on the vertical distances from the candidate outlier points to each fitting straight line to obtain the target outlier point, including: For any candidate outlier point, determine the minimum vertical distance from any candidate outlier point to each of the fitted straight lines; If the minimum vertical distance is greater than or equal to a preset threshold, any of the candidate abnormal points is determined to be the target abnormal point.

[0061] Specifically, when screening candidate outliers, the same operation is performed for each candidate outlier, that is, for any candidate outlier, the minimum vertical distance from the candidate outlier to each fitted line is determined. Then the size relationship between the minimum vertical distance and the preset threshold is determined. If the minimum vertical distance is greater than or equal to the preset threshold, it is considered that the candidate outlier is not on the fitted line corresponding to the minimum vertical distance, and it cannot be regarded as a normal point, and needs to be treated as a target outlier for subsequent removal. If the minimum vertical distance is less than the preset threshold, it is considered that the candidate outlier is on the fitted line corresponding to the minimum vertical distance, and it can be regarded as a normal point, and cannot be used as a target outlier.

[0062] like Figure 2 As shown above, in summary, the complete process of the laser radar point cloud outlier removal method provided in the embodiment of the present invention may include: Get each frame of point cloud data of the laser radar; Traverse each frame of point cloud data, determine the obstacle distance in the current frame of point cloud data, and extract the obstacle edge points in the current frame of point cloud data; According to the obstacle distance, determine the edge outlier points adjacent to the obstacle edge points; Determine candidate outlier points based on obstacle edge points and edge outlier points; Perform straight line fitting on the point cloud data of the current frame to obtain multiple fitting straight lines; Based on the vertical distance from the candidate outlier point to each fitted straight line, the minimum vertical distance from the candidate outlier point to each fitted straight line is determined, and it is judged whether the minimum vertical distance is greater than or equal to a preset threshold. If it is greater, it is determined as a target outlier point. If it is less than the preset threshold, the candidate outlier point is removed.

[0063] like Figure 3As shown, the laser radar 1 obtains the current frame point cloud data by scanning the obstacle 2 and the background 3, which contains a total of 500 laser radar points. If the number of the obstacle edge point is 0, the number of the laser radar point on the left side of the obstacle edge point is 499.

[0064] The target outlier points determined by the laser radar point cloud outlier point removal method provided in the embodiment of the present invention are as follows: Figure 4 The point cloud data of the current frame obtained after removing the circled points is as follows Figure 5 shown.

[0065] Among them, the obstacle distance d1 is 0.3m, k is calculated and rounded to 3, and the left and right sides of the obstacle are calculated to contain 7 candidate outliers. However, after the straight line fitting test, each obstacle outlier has a candidate outlier point on one side that is on a fitting line, so it cannot be used as a target outlier point. Finally, there are 6 target outliers in total.

[0066] from Figure 5 It can be seen that the point cloud data of the current frame after removal is clearer and more accurate. The trailing points and abnormal points on the edge of the obstacle are effectively removed, while the normal points are retained, avoiding the erroneous removal of valid data points. This helps to improve the accuracy and reliability of LiDAR data and provides better data support for subsequent map construction, obstacle detection and other tasks.

[0067] like Figure 6 As shown, based on the above embodiment, an embodiment of the present invention provides a laser radar point cloud abnormal point removal device, comprising: The data acquisition module 61 is used to acquire each frame of point cloud data of the laser radar; The outlier removal module 62 is used to traverse the point cloud data of each frame, determine the obstacle distance in the current frame point cloud data, determine the candidate outlier points of the obstacle edge in the current frame point cloud data based on the obstacle distance, determine the target outlier points based on the candidate outlier points, and remove the target outlier points.

[0068] On the basis of the above-mentioned embodiment, in the laser radar point cloud outlier removal device provided in the embodiment of the present invention, the outlier removal module is specifically used for: Extracting obstacle edge points in the current frame point cloud data; Based on the obstacle distance, determining an edge outlier point adjacent to the obstacle edge point; The candidate outlier point is determined based on the obstacle edge point and the edge outlier point.

[0069] On the basis of the above-mentioned embodiment, in the laser radar point cloud outlier removal device provided in the embodiment of the present invention, the outlier removal module is specifically used for: Determine a target distance range within which the obstacle distance lies; Based on the correspondence between the distance range and the number of edge outliers, determine the number of target edge outliers corresponding to the target distance range; Determining the edge outliers based on the number of the target edge outliers; Among them, in the corresponding relationship, a distance range with a large value corresponds to a small number of edge outliers.

[0070] On the basis of the above-mentioned embodiment, in the laser radar point cloud outlier removal device provided in the embodiment of the present invention, the corresponding relationship includes a first distance range, a second distance range and a third distance range whose values ​​increase in sequence; The number of edge outliers corresponding to the first distance range is greater than the number of edge outliers corresponding to the third distance range, and each value within the second distance range is inversely proportional to the corresponding number of edge outliers.

[0071] On the basis of the above-mentioned embodiment, in the laser radar point cloud outlier removal device provided in the embodiment of the present invention, the outlier removal module is further specifically used for: Performing straight line fitting on the point cloud data of the current frame to obtain a plurality of fitting straight lines; Based on the vertical distances from the candidate outlier points to each fitting straight line, the candidate outlier points are screened to obtain the target outlier points.

[0072] On the basis of the above-mentioned embodiment, in the laser radar point cloud outlier removal device provided in the embodiment of the present invention, the outlier removal module is further specifically used for: For any candidate outlier point, determine the minimum vertical distance from any candidate outlier point to each of the fitted straight lines; If the minimum vertical distance is greater than or equal to a preset threshold, any of the candidate abnormal points is determined to be the target abnormal point.

[0073] Specifically, the functions of each module in the laser radar point cloud outlier removal device provided in the embodiment of the present invention correspond one-to-one to the operating procedures of each step in the above-mentioned method embodiment, and the effects achieved are also consistent. Please refer to the above-mentioned embodiment for details, and no further details will be given in the embodiment of the present invention.

[0074] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7As shown, the electronic device may include: a laser radar 700, a processor 710, a communications interface 720, a memory 730 and a communication bus 740, wherein the laser radar 700 is used to collect each frame of point cloud data, and the laser radar 700, the processor 710, the communications interface 720 and the memory 730 can communicate with each other through the communication bus 740. The processor 710 can call the logic instructions in the memory 730 to execute the laser radar point cloud abnormal point removal method provided in the above embodiments.

[0075] It is understandable that the electronic device may be a robot, for example, a sweeping robot or other automatically moving intelligent robot, which is not specifically limited here.

[0076] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the relevant technology or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0077] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the laser radar point cloud outlier removal method provided by the above-mentioned methods.

[0078] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the laser radar point cloud outlier removal method provided by the above-mentioned methods.

[0079] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0080] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for removing abnormal points from a laser radar point cloud, characterized in that: include: Get each frame of point cloud data of the laser radar; Traversing each frame of point cloud data, determining the obstacle distance in the current frame of point cloud data, determining candidate outlier points at the edge of the obstacle in the current frame of point cloud data based on the obstacle distance, determining target outlier points based on the candidate outlier points, and removing the target outlier points.

2. The method for removing abnormal points from laser radar point cloud according to claim 1, characterized in that: The determining, based on the obstacle distance, candidate abnormal points at the obstacle edge in the current frame point cloud data includes: Extracting obstacle edge points in the current frame point cloud data; Based on the obstacle distance, determining an edge outlier point adjacent to the obstacle edge point; The candidate outlier point is determined based on the obstacle edge point and the edge outlier point.

3. The method for removing abnormal points from laser radar point cloud according to claim 2, characterized in that: The determining, based on the obstacle distance, an edge outlier point adjacent to the obstacle edge point comprises: Determine a target distance range within which the obstacle distance lies; Based on the correspondence between the distance range and the number of edge outliers, determine the number of target edge outliers corresponding to the target distance range; Determining the edge outliers based on the number of the target edge outliers; Among them, in the corresponding relationship, a distance range with a large value corresponds to a small number of edge outliers.

4. The method for removing abnormal points from laser radar point cloud according to claim 3, characterized in that: The corresponding relationship includes a first distance range, a second distance range and a third distance range whose values ​​increase in sequence; The number of edge outliers corresponding to the first distance range is greater than the number of edge outliers corresponding to the third distance range, and each value within the second distance range is inversely proportional to the corresponding number of edge outliers.

5. The method for removing abnormal points from laser radar point cloud according to any one of claims 1 to 4, characterized in that: The step of determining a target outlier point based on the candidate outlier point comprises: Performing straight line fitting on the point cloud data of the current frame to obtain a plurality of fitting straight lines; Based on the vertical distances from the candidate outlier points to each fitting straight line, the candidate outlier points are screened to obtain the target outlier points.

6. The method for removing abnormal points from laser radar point cloud according to claim 5, characterized in that: The step of screening the candidate outliers based on the vertical distances from the candidate outliers to each fitting straight line to obtain the target outliers includes: For any candidate outlier point, determine the minimum vertical distance from any candidate outlier point to each of the fitted straight lines; If the minimum vertical distance is greater than or equal to a preset threshold, any of the candidate abnormal points is determined to be the target abnormal point.

7. A laser radar point cloud outlier removal device, characterized in that: include: A data acquisition module is used to obtain point cloud data of each frame of the laser radar; The outlier removal module is used to traverse the point cloud data of each frame, determine the obstacle distance in the current frame point cloud data, determine the candidate outlier points of the obstacle edge in the current frame point cloud data based on the obstacle distance, determine the target outlier points based on the candidate outlier points, and remove the target outlier points.

8. An electronic device comprising a laser radar, a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The laser radar is used to collect each frame of point cloud data, and when the processor executes the program, it implements the laser radar point cloud outlier removal method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for removing abnormal points from a laser radar point cloud as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for removing abnormal points from a laser radar point cloud as described in any one of claims 1 to 6 is implemented.

Images