Point cloud matching detection method, device, equipment and storage medium

Abstract

The present invention relates to the field of image processing technology, and discloses a point cloud matching detection method, apparatus, device, and storage medium. The method comprises: obtaining multiple groups of point cloud data with matching relationships, and dividing each group of point cloud data into corresponding multiple grids; calculating the point cloud description information corresponding to each grid, and calculating the grid similarity between each group of point cloud data with matching relationships based on the point cloud description information; based on the grid similarity, determining whether the similarity ratio of the grids between each group of point cloud data meets a preset ratio threshold; if the similarity ratio meets the preset ratio threshold, fusing the grids of the corresponding group of point cloud data according to a preset scale to obtain a fusion result; based on the fusion result, evaluating the matching degree of the matching relationship between each group of point cloud data. The present invention improves the accuracy of the evaluation of the point cloud matching degree.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a point cloud matching detection method, device, equipment and storage medium. Background Art

[0002] Autonomous driving technology encompasses high-precision mapping, environmental perception, behavioral decision-making, path planning, motion control, and other technologies. With the rise of intelligent vehicles, autonomous driving holds broad application prospects. High-precision maps, the three-dimensional data upon which most autonomous driving technologies rely, are generated by processing point cloud data and positioning data using sensor fusion algorithms. During the high-precision map production process, multiple overlapping 3D point cloud data sets must be aligned within the same 3D coordinate system to form an accurate and complete model. Due to the massive amount of collected 3D data, the matching results of different 3D point cloud data sets during the point cloud mosaic production process directly impact the quality of the final point cloud mosaic.

[0003] Currently, high-precision maps are generated by extracting feature data from collected image data to reconstruct a macroscopic road model, thereby obtaining the approximate location and direction of the road. This method works well for feature-rich urban roads. However, the model processing effect is poor for scenes with simple and repetitive features, such as tunnels and elevated roads. This is especially true for scenes like those inside tunnels and under elevated roads, where positioning signals are poor, convergence is poor, and observations are insufficient. The high-precision map constructed from 3D point cloud matching results is difficult to meet the required requirements. This means that existing methods struggle to maintain accuracy in point cloud fusion detection for various scenarios. Summary of the Invention

[0004] The main purpose of the present invention is to solve the technical problem that existing methods are difficult to maintain accuracy in point cloud fusion detection of various scenes.

[0005] A first aspect of the present invention provides a point cloud matching detection method, comprising: obtaining multiple groups of point cloud data with matching relationships, and dividing each group of point cloud data into corresponding multiple grids; calculating the point cloud description information corresponding to each grid, and calculating the grid similarity between each group of point cloud data with matching relationships based on the point cloud description information; based on the grid similarity, judging whether the similarity ratio of the grids between each group of point cloud data meets a preset ratio threshold; if the similarity ratio meets the preset ratio threshold, fusing the grids of the corresponding group of point cloud data according to a preset scale to obtain a fusion result; based on the fusion result, evaluating the matching degree of the matching relationship between each group of point cloud data.

[0006] Optionally, in a first implementation method of the first aspect of the present invention, the calculation of the point cloud description information corresponding to each grid includes: determining the position information of the point cloud contained in each grid, and based on the position information, respectively calculating the position mean of the point cloud contained in each grid; based on the point cloud position mean and the position information, respectively calculating the position covariance of the point cloud contained in each grid; and obtaining the point cloud description information corresponding to each grid according to the position mean and the position covariance.

[0007] Optionally, in a second implementation method of the first aspect of the present invention, the grid similarity between each group of point cloud data with a matching relationship is calculated based on the point cloud description information, including: calculating the mean of the position covariance between the grids with a matching relationship between each group of point cloud data, and calculating the difference between the position means between the grids with a matching relationship between each group of point cloud data; based on the position of the point cloud data of the group in which each grid is located, constructing a first matrix corresponding to the mean of the position covariance, and constructing a second matrix corresponding to the difference of the position means; based on the first matrix and the second matrix, calculating the grid distance between each group of point cloud data with a matching relationship, and using it as the corresponding grid similarity.

[0008] Optionally, in a third implementation method of the first aspect of the present invention, if the similarity ratio satisfies a preset ratio threshold, the grids of the corresponding group of point cloud data are fused according to a preset scale to obtain a fusion result, including: if the similarity ratio satisfies a preset ratio threshold, the grids of the corresponding group of point cloud data are adjacently merged according to a grid of a preset scale to obtain a fused grid; determining a point cloud description information combination corresponding to each fused grid, and performing superposition calculation on each point cloud description information combination according to a preset algorithm to obtain fused point cloud description information; updating the fused point cloud description information to the fused grid to obtain a fusion result.

[0009] Optionally, in a fourth implementation method of the first aspect of the present invention, if the similarity ratio satisfies a preset ratio threshold, each grid of the corresponding group point cloud data is adjacently merged according to a grid of a preset scale to obtain a fused grid, including: if the similarity ratio satisfies a preset ratio threshold, at least one layer of adjacent grids of each grid of the corresponding group point cloud data is determined according to a preset scale; each grid is merged with the corresponding at least one layer of adjacent grids to obtain a fused grid.

[0010] Optionally, in a fifth implementation manner of the first aspect of the present invention, the matching degree of the matching relationship between each group of point cloud data is evaluated based on the fusion result, including: based on the fusion result, judging whether the size of the fused grid meets a preset size threshold; if the size of the fused grid meets the preset size threshold, determining whether the matching degree of the matching relationship between the corresponding groups of point cloud data meets the preset quality requirements; if the scale of the fused grid does not meet the preset scale threshold, judging whether the similarity ratio of the fused grid in the corresponding group of point cloud data meets the preset threshold based on the fused point cloud description information; if the similarity ratio of the fused grid does not meet the preset threshold, determining that the matching degree of the matching relationship between the corresponding groups of point cloud data does not meet the preset quality requirements.

[0011] The second aspect of the present invention provides a point cloud matching detection device, comprising: a division module for acquiring multiple groups of point cloud data with matching relationships, and dividing each group of point cloud data into corresponding multiple grids; a similarity calculation module for calculating the point cloud description information corresponding to each grid, and calculating the grid similarity between each group of point cloud data with matching relationships based on the point cloud description information; a proportion judgment module for judging whether the similarity ratio of the grids between each group of point cloud data meets a preset ratio threshold based on the grid similarity; a fusion module for fusing the grids of the corresponding group of point cloud data according to a preset scale to obtain a fusion result if the similarity ratio meets the preset ratio threshold; and an evaluation module for evaluating the matching degree of the matching relationship between each group of point cloud data based on the fusion result.

[0012] Optionally, in a first implementation method of the second aspect of the present invention, the similarity calculation module includes: a mean calculation unit, used to determine the position information of the point cloud contained in each grid, and based on the position information, respectively calculate the position mean of the point cloud contained in each grid; a covariance calculation unit, used to respectively calculate the position covariance of the point cloud contained in each grid based on the point cloud position mean and the position information; a production unit, used to obtain the point cloud description information corresponding to each grid based on the position mean and the position covariance.

[0013] Optionally, in a second implementation of the second aspect of the present invention, the similarity calculation module further includes: a difference calculation unit, used to calculate the mean of the position covariance between the grids that have a matching relationship between each group of point cloud data, and to calculate the difference of the position means between the grids that have a matching relationship between each group of point cloud data; a construction unit, used to construct a first matrix corresponding to the mean of the position covariance based on the position of the point cloud data of the group to which each grid is located, and to construct a second matrix corresponding to the difference of the position means; a distance calculation unit, used to calculate the distance between the grids that have a matching relationship between each group of point cloud data based on the first matrix and the second matrix, and use it as the corresponding grid similarity.

[0014] Optionally, in a third implementation of the second aspect of the present invention, the fusion module includes: a merging unit, which is used to perform adjacent merging of each grid of the corresponding group point cloud data according to a preset scale grid if the similarity ratio meets a preset ratio threshold, to obtain a fused grid; an overlay calculation unit, which is used to determine the point cloud description information combination corresponding to each fused grid, and perform overlay calculation on each point cloud description information combination according to a preset algorithm to obtain fused point cloud description information; and an updating unit, which is used to update the fused point cloud description information to the fused grid to obtain a fusion result.

[0015] Optionally, in a fourth implementation method of the second aspect of the present invention, the merging unit is further used to: if the similarity ratio satisfies a preset ratio threshold, determine at least one layer of adjacent grids for each grid of the corresponding group point cloud data according to a preset scale; merge each grid with the corresponding at least one layer of adjacent grids to obtain a fused grid.

[0016] Optionally, in a fifth implementation of the second aspect of the present invention, the evaluation module includes: a size determination unit, for determining, based on the fusion result, whether the size of the fused grid meets a preset size threshold; a quality evaluation unit, for determining, if the size of the fused grid meets the preset size threshold, whether the degree of matching of the matching relationship between the corresponding groups of point cloud data meets the preset quality requirement; a ratio determination unit, for determining, based on the fused point cloud description information, whether the similarity ratio of the fused grid in the corresponding group of point cloud data meets the preset threshold if the scale of the fused grid does not meet the preset scale threshold; the quality evaluation unit is also for determining, if the similarity ratio of the fused grid does not meet the preset threshold, that the degree of matching of the matching relationship between the corresponding groups of point cloud data does not meet the preset quality requirement.

[0017] The third aspect of the present invention provides a point cloud matching detection device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the point cloud matching detection device executes the above-mentioned point cloud matching detection method.

[0018] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, enable the computer to execute the above-mentioned point cloud matching detection method.

[0019] The technical solution provided by the present invention obtains multiple sets of matching point cloud data and divides each set of point cloud data into corresponding multiple grids. The point cloud description information corresponding to each grid is calculated, and based on the point cloud description information, the grid similarity between each set of matching point cloud data is calculated. Based on the grid similarity, it is determined whether the similarity ratio of the grids between each set of point cloud data meets a preset ratio threshold. If the similarity ratio meets the preset ratio threshold, the grids of the corresponding set of point cloud data are fused according to a preset scale to obtain a fusion result. Based on the fusion result, the matching degree of each set of point cloud data is evaluated. The present invention improves the accuracy of the assessment of the matching degree of point clouds. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Schematic diagram of a first embodiment of a point cloud matching detection method according to an embodiment of the present invention;

[0021] Figure 2 Schematic diagram of a second embodiment of a point cloud matching detection method according to an embodiment of the present invention;

[0022] Figure 3 Schematic diagram of an embodiment of a detection device for point cloud matching in an embodiment of the present invention;

[0023] Figure 4 Schematic diagram of another embodiment of a detection device for point cloud matching in an embodiment of the present invention;

[0024] Figure 5 Schematic diagram of an embodiment of a point cloud matching detection device in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] Embodiments of the present invention provide a point cloud matching detection method, apparatus, device, and storage medium. These methods acquire multiple sets of point cloud data with matching relationships and divide each set of point cloud data into corresponding multiple grids. The method calculates point cloud description information corresponding to each grid, and based on the point cloud description information, calculates the grid similarity between each set of point cloud data with matching relationships. Based on the grid similarity, the method determines whether the similarity ratio of the grids between each set of point cloud data meets a preset ratio threshold. If the similarity ratio meets the preset ratio threshold, the grids of the corresponding set of point cloud data are fused according to a preset scale to obtain a fusion result. Based on the fusion result, the matching degree of the matching relationship between each set of point cloud data is evaluated. This invention improves the accuracy of point cloud matching assessment.

[0026] The terms "first," "second," "third," "fourth," and so on (if any) in the description and claims of the present invention and in the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that shown or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus that includes a series of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, product, or apparatus.

[0027] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 The first embodiment of the point cloud matching detection method in the embodiment of the present invention includes:

[0028] 101. Acquire multiple sets of point cloud data having a matching relationship, and divide each set of point cloud data into corresponding multiple grids;

[0029] It is understandable that the execution subject of the present invention can be a detection device for point cloud matching, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking the server as the execution subject as an example.

[0030] In this embodiment, multiple point cloud lidars are configured on the autonomous vehicle / conventional vehicle to collect scene point cloud information during driving. The multiple sets of point cloud data with matching relationships include: multiple sets of point cloud data collected by different radars with partial scene overlap; and multiple sets of point cloud data collected by the same radar in different frames with partial scene overlap (for ease of explanation, the following description uses two sets of point cloud data). The two sets of point cloud data here refer to point cloud data that has been registered. Therefore, after the two sets of point cloud data are divided into multiple grids, corresponding matching relationships also exist between the grids.

[0031] Specifically, for example, if the registration results of two sets of point cloud data show that 30% of the point cloud areas are overlapping areas, then after being divided into multiple grids, the grids in the 30% overlapping areas of each set of point cloud data will also have corresponding matching relationships. Subsequently, it is necessary to detect whether this part of the overlapping areas is consistent to detect whether the grids of the two sets of point cloud data are similar; and to detect whether the non-overlapping parts are inconsistent to detect whether the grids of the two sets of point cloud data are dissimilar.

[0032] 102. Calculate the point cloud description information corresponding to each grid, and calculate the grid similarity between each set of point cloud data with a matching relationship based on the point cloud description information;

[0033] In this embodiment, the feature description of each mesh is represented by calculating its point cloud description information. This point cloud description information can include local feature descriptors such as point feature histograms (PFH) and fast point feature histograms (FPFH), as well as statistics such as the point cloud mean, variance, and covariance. By quantifying the point cloud feature description of the meshes, subsequent similarity comparisons of the meshes are facilitated.

[0034] In this embodiment, after the point cloud description information of each grid is obtained by quantitative calculation, the matching relationship of the grids between the two sets of point cloud data is compared one by one to determine the similarity between each two matching grids, so as to measure whether the point cloud data of the same scene area displayed by the two sets of point cloud data are consistent, and whether the point cloud data of different scene areas are inconsistent.

[0035] 103. Based on the grid similarity, determine whether the grid similarity ratio between each set of point cloud data meets a preset ratio threshold;

[0036] In this embodiment, the ratio threshold here refers to whether the similarity of each corresponding grid between the two sets of point cloud data exceeds the preset similarity threshold, so that the similarity grid ratio threshold between the corresponding grids in the two sets of point cloud data can be calculated; by comparing the similarity of each grid between the two sets of point cloud data one by one, it is determined whether the grid similarity ratio corresponding to each grid between the two sets of point cloud data is greater than or equal to the preset ratio threshold.

[0037] In this embodiment, by comparing the grid similarity between the two sets of point cloud data one by one, it is possible to determine whether the grids between the two are similar, thereby determining the similar grids at the relevant scale, so that the point cloud data that does not meet the conditions can be subsequently scale-fused to evaluate the degree of matching between the current two point cloud data.

[0038] 104. If the similarity ratio meets the preset ratio threshold, each grid of the point cloud data of the corresponding group is fused according to the preset scale to obtain a fusion result;

[0039] In this embodiment, if the similarity ratios of the grids between two groups of point cloud data meet a preset ratio threshold, grid fusion is performed on the grids of the corresponding groups of point cloud data according to a preset scale to obtain a fusion result.

[0040] In this embodiment, each grid of the point cloud data of the corresponding group is grid-fused according to a preset scale by multiplying the grid corresponding to the current point cloud data, and then recalculating the similarity and judging the similarity threshold ratio of the multiplied point cloud data to obtain the result of the grid fusion corresponding to the point cloud data.

[0041] 105. Based on the fusion results, evaluate the matching degree of the matching relationship between each set of point cloud data.

[0042] In this embodiment, based on the results of the above fusion, the matching degree between the matching relationships between each group of corresponding point cloud data is evaluated, thereby evaluating the quality of the current algorithm for matching the point cloud data. This not only achieves better processing of urban roads with rich features, but also can accurately and quickly evaluate the point cloud matching results in scenes such as tunnels, elevated tunnels, and under elevated roads with single and repetitive features, poor GPS signals, poor convergence processing, and insufficient observations, thereby improving the robustness of the point cloud matching result evaluation for different scenes.

[0043] In an embodiment of the present invention, multiple sets of matching point cloud data are obtained, and each set of point cloud data is divided into multiple corresponding grids. The point cloud description information corresponding to each grid is calculated, and based on the point cloud description information, the grid similarity between each set of matching point cloud data is calculated. Based on the grid similarity, it is determined whether the similarity ratio of the grids between each set of point cloud data meets a preset ratio threshold. If the similarity ratio meets the preset ratio threshold, the grids of the corresponding set of point cloud data are fused according to a preset scale to obtain a fusion result. Based on the fusion result, the matching degree of the matching relationship between each set of point cloud data is evaluated. This invention improves the accuracy of the assessment of the matching degree of point clouds.

[0044] See also Figure 2 A second embodiment of the point cloud matching detection method in the embodiment of the present invention includes:

[0045] 201. Acquire multiple sets of point cloud data having matching relationships, and divide each set of point cloud data into corresponding multiple grids;

[0046] In this embodiment, multiple groups of point cloud data with matching relationships matched by the algorithm to be evaluated are obtained, and each group of point cloud data is divided into voxel grids of corresponding sizes. For example, the matched point cloud A and point cloud B are divided into voxel grids according to a certain size (for example, 25 cm).

[0047] 202. Determine the position information of the point cloud contained in each grid, and calculate the position mean of the point cloud contained in each grid based on the position information;

[0048] In this embodiment, the position information here refers to the position information corresponding to multiple point clouds contained in each grid (voxel grid), and each grid contains at least one point cloud. The position mean here refers to the mean of the positions of each voxel grid in the entire point cloud data. The calculation formula for the position mean here is , where n represents the number of corresponding point clouds, position i Indicates the position information of the nth point cloud, and mean is the mean.

[0049] In practical applications, by determining the position information group of the point cloud contained in each corresponding grid in the two groups, and then based on the position information corresponding to each grid, the position mean information of the corresponding voxel grid contained in each point cloud data is calculated respectively.

[0050] 203. Based on the point cloud position mean and position information, respectively calculate the position covariance of the point cloud contained in each grid;

[0051] In this embodiment, the position covariance (cov) here is calculated by the point cloud mean and position information, and the corresponding calculation formula is: , where cov is the position covariance.

[0052] In practical applications, the above-mentioned point cloud position mean and position information are substituted into the position covariance formula to calculate the position covariance value of the point cloud contained in each grid corresponding to each set of point cloud data.

[0053] 204. Obtaining point cloud description information corresponding to each grid according to the position mean and position covariance, and calculating the mean of the position covariance between the grids having a matching relationship between each set of point cloud data, and calculating the difference of the position mean between the grids having a matching relationship between each set of point cloud data;

[0054] In this embodiment, based on the above-mentioned position mean and position covariance, the point cloud description information corresponding to each grid is obtained, and then through mean calculation, the mean of the position covariance between the grids with matching relationships between each group of point cloud data is calculated, and the difference in the position mean between the grids with matching relationships between each group of point cloud data is calculated.

[0055] In this embodiment, before performing grid similarity judgment, it is first determined whether the current voxel grid exceeds the preset grid size threshold. In this case, it can be directly considered that each group of point clouds is similar at all scales. If it does not exceed, after calculating the point cloud descriptor in each grid, the grid similarity of the grid corresponding to each group of point clouds is required.

[0056] 205. Based on the position of the point cloud data of the group to which each grid belongs, construct a first matrix corresponding to the mean of the position covariance, and construct a second matrix corresponding to the difference of the position means;

[0057] In this embodiment, the formula for the mean of the position covariance is: , where cov a and cov b is the covariance of the voxel grid corresponding to each group of point clouds, and cov is the mean of the position covariance of the corresponding voxel grid; the formula of the first matrix here is , where mean a and mean b is the mean of the voxel grid corresponding to each group of point clouds, and T is the transpose; the formula of the second matrix here is , det is the determinant.

[0058] In practical applications, based on the position of the point cloud data of each grid group, the formula of the mean of the position covariance is first used to construct the mean of the position covariance of the voxel grid corresponding to each group of point clouds, and then the first matrix is ​​calculated using the mean of the position covariance of the voxel grid corresponding to each group of point clouds and the formula of the first matrix, and the second matrix corresponding to the difference of the position means is constructed using the formula of the second matrix.

[0059] 206. Based on the first matrix and the second matrix, calculate the grid distance between each set of point cloud data that has a matching relationship, and use it as the corresponding grid similarity;

[0060] In this embodiment, based on the first matrix and the second matrix, the first matrix and the second matrix are added together to calculate the grid distance between each set of point cloud data having a matching relationship, and use it as the corresponding grid similarity.

[0061] 207. Based on the grid similarity, determine whether the grid similarity ratio between each set of point cloud data meets a preset ratio threshold;

[0062] In this embodiment, based on the calculated mesh similarity of each group of point clouds, it is determined whether the similarity ratio of the meshes between each group of point cloud data meets a preset ratio threshold.

[0063] 208. If the similarity ratio meets a preset ratio threshold, each grid of the corresponding group of point cloud data is adjacently merged according to a preset scale grid to obtain a fused grid;

[0064] In this embodiment, the grid fusion calculation formula here is , ,in , num k is the number of fusions, num i is the number of fused voxel grids, mean is the mean before fusion, cov is the covariance before fusion, cov i is the covariance value corresponding to the nth fused voxel grid.

[0065] In this embodiment, if the similarity ratio meets a preset ratio threshold, the voxel grid corresponding to each point cloud group is first multiplied (for example, the current voxel grid is 25 cm, and after multiplication it is 50 cm). The multiplied voxel grid is then scale-fused with the original corresponding point cloud descriptor. Specifically, the point cloud description information combination corresponding to each fused grid is determined, and each point cloud description information combination is superimposed and calculated according to a preset algorithm. At least one layer of adjacent grids for each grid in the corresponding group of point cloud data is determined according to a preset scale (e.g., 2*2*2, a total of 8 adjacent grids). Each grid is then merged with the corresponding at least one layer of adjacent grids to obtain a fused grid. Furthermore, if the similarity ratio does not meet the preset ratio threshold, a determination is made as to whether the proportion of the currently dissimilar grid in the entire voxel grid exceeds a preset proportion (e.g., 60%). If so, the degree of matching between each point cloud data group is directly evaluated. If not, scale fusion is performed and the similarity determination is repeated. If the conditions are met, the degree of matching between each point cloud data group is evaluated.

[0066] 209. Determine the point cloud description information combination corresponding to each fused grid, and perform superposition calculation on each point cloud description information combination according to a preset algorithm to obtain fused point cloud description information;

[0067] In this embodiment, by determining the point cloud description information combination corresponding to each fused grid, each point cloud description information combination is superimposed and calculated according to a preset algorithm, and the newly calculated covariance corresponding to each voxel grid is superimposed with the covariance before fusion, as well as the newly calculated mean and the mean before fusion, to obtain the fused point cloud description information.

[0068] 210. Update the fused point cloud description information to the fused grid to obtain a fusion result, and based on the fusion result, evaluate the matching degree of the matching relationship between each set of point cloud data.

[0069] In this embodiment, the fusion result is obtained by updating the fused point cloud description information to the fused grid. Based on the fusion result, it is determined whether the size of the fused grid meets the preset size threshold; if the size of the fused grid meets the preset size threshold, it is determined that the matching degree of the matching relationship between the corresponding groups of point cloud data meets the preset quality requirement; if the scale of the fused grid does not meet the preset scale threshold, based on the fused point cloud description information, it is determined whether the similarity ratio of the fused grid in the corresponding group of point cloud data meets the preset threshold; if the similarity ratio of the fused grid does not meet the preset threshold, it is determined that the matching degree of the matching relationship between the corresponding groups of point cloud data does not meet the preset quality requirement.

[0070] Acquire multiple sets of matching point cloud data and divide each set of point cloud data into corresponding grids; calculate the point cloud description information corresponding to each grid, and based on the point cloud description information, calculate the grid similarity between each set of matching point cloud data; based on the grid similarity, determine whether the similarity ratio of the grids between each set of point cloud data meets a preset ratio threshold; if the similarity ratio meets the preset ratio threshold, fuse the grids of the corresponding set of point cloud data according to a preset scale to obtain a fusion result; based on the fusion result, evaluate the matching degree of the matching relationship between each set of point cloud data. This invention improves the accuracy of the assessment of the matching degree of point clouds.

[0071] The above describes the detection method of point cloud matching in the embodiment of the present invention. The following describes the detection device of point cloud matching in the embodiment of the present invention. Figure 3 In one embodiment of the present invention, a detection device for point cloud matching includes:

[0072] A division module 301 is used to obtain multiple groups of point cloud data with matching relationships, and to divide each group of point cloud data into corresponding multiple grids;

[0073] Similarity calculation module 302, used to calculate the point cloud description information corresponding to each grid, and calculate the grid similarity between each set of point cloud data with a matching relationship based on the point cloud description information;

[0074] The ratio determination module 303 is used to determine whether the similarity ratio of the grids between each set of point cloud data meets a preset ratio threshold based on the grid similarity;

[0075] The fusion module 304 is configured to fuse the grids of the corresponding group of point cloud data according to a preset scale to obtain a fusion result if the similarity ratio meets a preset ratio threshold;

[0076] The evaluation module 305 is used to evaluate the matching degree of the matching relationship between each set of point cloud data based on the fusion result.

[0077] In an embodiment of the present invention, multiple groups of point cloud data with matching relationships are obtained, and each group of point cloud data is divided into corresponding multiple grids; the point cloud description information corresponding to each grid is calculated, and based on the point cloud description information, the grid similarity between each group of point cloud data with matching relationships is calculated; based on the grid similarity, it is determined whether the similarity ratio of the grids between each group of point cloud data meets a preset ratio threshold; if the similarity ratio meets the preset ratio threshold, the grids of the corresponding group of point cloud data are fused according to a preset scale to obtain a fusion result; based on the fusion result, the matching degree of the matching relationship between each group of point cloud data is evaluated. The present invention covers scenes such as urban roads with dense features and tunnels with fewer features by fusing from a small grid scale to a large scale, and has strong adaptability to scene changes. In addition, for certain special descriptors, the complex calculation only needs to be performed once. Subsequently, when performing grid fusion, adjacent grids only need to obtain the descriptor corresponding to the large-scale grid through simple calculation, thereby improving the accuracy of the assessment of the point cloud matching degree.

[0078] See also Figure 4 Another embodiment of the point cloud matching detection device in the embodiment of the present invention includes:

[0079] A division module 301 is used to obtain multiple groups of point cloud data with matching relationships, and to divide each group of point cloud data into corresponding multiple grids;

[0080] A similarity calculation module 302 is used to calculate the point cloud description information corresponding to each grid, and calculate the grid similarity between each set of point cloud data having a matching relationship based on the point cloud description information;

[0081] A ratio determination module 303 is configured to determine, based on the grid similarity, whether the similarity ratio of the grids between each set of point cloud data meets a preset ratio threshold;

[0082] A fusion module 304 is configured to fuse the grids of the corresponding group of point cloud data according to a preset scale to obtain a fusion result if the similarity ratio meets a preset ratio threshold;

[0083] The evaluation module 305 is used to evaluate the matching degree of the matching relationship between each set of point cloud data based on the fusion result.

[0084] Specifically, the similarity calculation module 302 includes:

[0085] The mean calculation unit 3021 is used to determine the position information of the point cloud contained in each grid, and based on the position information, calculate the position mean of the point cloud contained in each grid;

[0086] A covariance calculation unit 3022 is configured to calculate the position covariance of the point cloud contained in each grid based on the point cloud position mean and the position information;

[0087] The production unit 3023 is used to obtain point cloud description information corresponding to each grid according to the position mean and the position covariance.

[0088] Specifically, the similarity calculation module 302 further includes:

[0089] a difference calculation unit 3024, configured to calculate the mean of the position covariances between the grids that have a matching relationship between each set of point cloud data, and to calculate the difference of the position means between the grids that have a matching relationship between each set of point cloud data;

[0090] A construction unit 3025 is configured to construct a first matrix corresponding to the mean of the position covariances and a second matrix corresponding to the differences of the position means based on the positions of the point cloud data of the group to which each grid belongs;

[0091] The distance calculation unit 3026 is used to calculate the grid distance between each set of point cloud data with a matching relationship based on the first matrix and the second matrix, and use it as the corresponding grid similarity.

[0092] Specifically, the fusion module 304 includes:

[0093] The merging unit 3041 is configured to perform neighbor merging on each grid of the corresponding group of point cloud data according to a preset scale to obtain a fused grid if the similarity ratio meets a preset ratio threshold;

[0094] The superposition calculation unit 3042 is used to determine the point cloud description information combination corresponding to each fused grid, and perform superposition calculation on each point cloud description information combination according to a preset algorithm to obtain the fused point cloud description information;

[0095] The updating unit 3043 is used to update the fused point cloud description information to the fused grid to obtain a fusion result.

[0096] Specifically, the merging unit 3041 is further configured to:

[0097] If the similarity ratio satisfies a preset ratio threshold, determining at least one layer of adjacent grids for each grid of the corresponding group of point cloud data according to a preset scale;

[0098] Each grid is merged with at least one layer of corresponding adjacent grids to obtain a fused grid.

[0099] Specifically, the evaluation module 305 includes:

[0100] A size determination unit 3051 is configured to determine, based on the fusion result, whether the size of the fused grid meets a preset size threshold;

[0101] The quality evaluation unit 3052 is configured to determine whether the degree of matching between the corresponding groups of point cloud data meets a preset quality requirement if the size of the fused grid meets a preset size threshold;

[0102] A ratio determination unit 3053 is configured to determine, based on the fused point cloud description information, whether a similarity ratio of the fused grids in the corresponding group of point cloud data meets a preset threshold if the scale of the fused grid does not meet a preset scale threshold;

[0103] The quality assessment unit 3052 is further configured to determine that the degree of matching between the corresponding groups of point cloud data does not meet a preset quality requirement if the similarity ratio of the fused grids does not meet a preset threshold.

[0104] above Figure 3 and Figure 4 The point cloud matching detection device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The point cloud matching detection device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0105] Figure 5 Figure 5 is a schematic diagram of the structure of a point cloud matching detection device provided by an embodiment of the present invention. This point cloud matching detection device 500 may vary significantly depending on configuration or performance. It may include one or more central processing units (CPUs) 510 (e.g., one or more processors), memory 520, and one or more storage media 530 (e.g., one or more mass storage devices) storing application programs 533 or data 532. The memory 520 and storage medium 530 may be either transient or persistent storage. The program stored in the storage medium 530 may include one or more modules (not shown), each of which may include a series of instruction operations within the point cloud matching detection device 500. Furthermore, the processor 510 may be configured to communicate with the storage medium 530 to execute the series of instruction operations stored in the storage medium 530 on the point cloud matching detection device 500.

[0106] The point cloud matching detection device 500 may further include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input and output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 5 The structure of the point cloud matching detection device shown does not constitute a limitation on the point cloud matching detection device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0107] The present invention also provides a point cloud matching detection device, wherein the computer device includes a memory and a processor, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs the steps of the point cloud matching detection method described in each of the above embodiments. The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions. When the instructions are executed on a computer, the computer performs the steps of the point cloud matching detection method.

[0108] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0109] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0110] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. 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 point cloud matching detection method, characterized in that: The point cloud matching detection method includes: Acquire multiple sets of point cloud data with matching relationships, and divide each set of point cloud data into corresponding multiple grids; Calculate the point cloud description information corresponding to each grid, and based on the point cloud description information, calculate the grid similarity between each set of point cloud data that has a matching relationship, so as to measure whether the point cloud data of the same scene area displayed by the two sets of point cloud data registration are consistent, and whether the point cloud data of different scene areas are inconsistent; Based on the grid similarity, determining whether the grid similarity ratio between each set of point cloud data meets a preset ratio threshold; If the similarity ratio meets a preset ratio threshold, each grid of the point cloud data of the corresponding group is fused according to a preset scale to obtain a fusion result; If the similarity ratio does not meet the preset ratio threshold, it is determined whether the proportion of the current dissimilar grid in the entire voxel grid exceeds the preset proportion value. If so, the matching degree of the matching relationship between each set of point cloud data is directly evaluated; if not, scale fusion is performed, wherein the scale is fused from the small grid scale to the large scale; Based on the fusion result, the matching degree of the matching relationship between each set of point cloud data is evaluated.

2. The point cloud matching detection method according to claim 1, characterized in that: The calculation of the point cloud description information corresponding to each grid includes: Determining position information of the point clouds contained in each grid, and calculating the position mean of the point clouds contained in each grid based on the position information; Calculating the position covariance of the point cloud contained in each grid based on the point cloud position mean and the position information; According to the position mean and the position covariance, point cloud description information corresponding to each grid is obtained.

3. The point cloud matching detection method according to claim 2, characterized in that: Calculating the grid similarity between each set of point cloud data in a matching relationship based on the point cloud description information includes: Calculate the mean of the position covariance between the grids that have a matching relationship between each set of point cloud data, and calculate the difference of the position mean between the grids that have a matching relationship between each set of point cloud data; Based on the position of the point cloud data of the group to which each grid belongs, constructing a first matrix corresponding to the mean of the position covariance, and constructing a second matrix corresponding to the difference of the position means; Based on the first matrix and the second matrix, a grid distance between each set of point cloud data having a matching relationship is calculated and used as the corresponding grid similarity.

4. The point cloud matching detection method according to claim 1, characterized in that: If the similarity ratio satisfies a preset ratio threshold, each grid of the corresponding group of point cloud data is fused according to a preset scale to obtain a fusion result, including: If the similarity ratio meets a preset ratio threshold, each grid of the corresponding group of point cloud data is adjacently merged according to a preset scale grid to obtain a fused grid; Determine the point cloud description information combination corresponding to each fused grid, and perform superposition calculation on each point cloud description information combination according to a preset algorithm to obtain the fused point cloud description information; The fused point cloud description information is updated to the fused grid to obtain a fusion result.

5. The point cloud matching detection method according to claim 4, characterized in that: If the similarity ratio satisfies a preset ratio threshold, each grid of the corresponding group of point cloud data is adjacently merged according to a preset scale to obtain a fused grid, including: If the similarity ratio satisfies a preset ratio threshold, determining at least one layer of adjacent grids for each grid of the corresponding group of point cloud data according to a preset scale; Each grid is merged with at least one layer of corresponding adjacent grids to obtain a fused grid.

6. The point cloud matching detection method according to claim 4, characterized in that: The step of evaluating the matching degree of the matching relationship between each set of point cloud data based on the fusion result includes: Based on the fusion result, determining whether the size of the fused grid meets a preset size threshold; If the size of the fused grid meets the preset size threshold, it is determined that the matching degree between the corresponding groups of point cloud data meets the preset quality requirements; If the scale of the fused grid does not meet the preset scale threshold, then based on the fused point cloud description information, it is determined whether the similarity ratio of the fused grid in the corresponding group of point cloud data meets the preset threshold; If the similarity ratio of the fused grids does not meet a preset threshold, it is determined that the degree of matching between the corresponding groups of point cloud data does not meet a preset quality requirement.

7. A point cloud matching detection device, characterized in that: The point cloud matching detection device includes: A division module is used to obtain multiple groups of point cloud data with matching relationships and divide each group of point cloud data into corresponding multiple grids; A similarity calculation module is used to calculate the point cloud description information corresponding to each grid, and based on the point cloud description information, calculate the grid similarity between each set of point cloud data with a matching relationship, so as to measure whether the point cloud data of the same scene area displayed by the registration of two sets of point cloud data are consistent, and whether the point cloud data of different scene areas are inconsistent; A ratio determination module is used to determine whether the similarity ratio of the grids between each set of point cloud data meets a preset ratio threshold based on the grid similarity; A fusion module is configured to fuse the grids of the corresponding group of point cloud data according to a preset scale to obtain a fusion result if the similarity ratio meets a preset ratio threshold; if the similarity ratio does not meet the preset ratio threshold, determine whether the proportion of the current dissimilar grid in the entire voxel grid exceeds a preset proportion value, and if not, perform scale fusion, wherein the fusion is performed from a small grid scale to a large scale; An evaluation module, configured to evaluate the degree of matching between each set of point cloud data based on the fusion result; The evaluation module is further configured to directly evaluate the degree of matching between each set of point cloud data when the similarity ratio does not meet a preset ratio threshold and the proportion of the current dissimilar grid in the entire voxel grid exceeds a preset proportion value.

8. The point cloud matching detection device according to claim 7, characterized in that: The similarity calculation module includes: a mean calculation unit, configured to determine position information of the point clouds contained in each grid, and based on the position information, respectively calculate the position mean of the point clouds contained in each grid; a covariance calculation unit, configured to calculate the position covariance of the point cloud contained in each grid based on the point cloud position mean and the position information; The production unit is used to obtain point cloud description information corresponding to each grid according to the position mean and the position covariance.

9. A point cloud matching detection device, characterized in that: The point cloud matching detection device includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instructions in the memory to enable the point cloud matching detection device to perform the steps of the point cloud matching detection method according to any one of claims 1 to 6.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the point cloud matching detection method as described in any one of claims 1 to 6 are implemented.

Images