A three-dimensional terrain mapping system and method based on laser radar

Abstract

The present application provides a laser radar-based three-dimensional terrain surveying and mapping system and method, which relates to the technical field of terrain surveying and mapping, and obtains laser radar point cloud data of a specified area; determines the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data, constructs an adaptive decision value according to the surveying and mapping features of all surveying and mapping points, performs edge screening on the surveying and mapping points according to the adaptive decision value, obtains a surveying and mapping edge point set, and determines a surveying and mapping feature descriptor of each surveying and mapping point through the surveying and mapping edge point set; performs incremental recovery on each surveying and mapping feature descriptor to obtain a corresponding coordinate increment, reconstructs the surveying and mapping point according to the corresponding coordinate increment, and obtains reconstructed laser radar point cloud data of the specified area; constructs a three-dimensional terrain model from the reconstructed laser radar point cloud data, and obtains three-dimensional terrain surveying and mapping data of the specified area based on the three-dimensional terrain model. The present application can effectively retain edge features when denoising the point cloud data, so as to improve the accuracy of the three-dimensional terrain model.

Description

Technical Field

[0001] The present application relates to the technical field of topographic surveying and mapping, and more specifically, to a three-dimensional topographic surveying and mapping system and method based on laser radar. Background Art

[0002] Three-dimensional terrain mapping based on LiDAR (Light Detection and Ranging) is a high-precision mapping technology. This technology is fast and efficient, can penetrate vegetation cover, and accurately capture surface details. It is widely used in urban planning, infrastructure construction, disaster monitoring, forest management and other fields. In operation, LiDAR can be carried on drones, helicopters, fixed-wing aircraft or ground platforms, and can be flexibly selected according to different application scenarios. Through point cloud processing, classification and modeling, a variety of terrain results such as high-resolution digital elevation models, digital surface models and orthophotos can be generated.

[0003] In the prior art, the volume of collected point cloud data is huge, and the processing of large-scale point cloud data involves multiple complex links, including noise removal, point cloud segmentation, classification, registration, interpolation modeling, etc. Each link poses a severe test to the performance of computer hardware and the efficiency of data processing algorithms. When using LiDAR to collect point cloud data, due to the instability of the system within the scanning process, the different reflectivity of the surface of the object to be measured, and other problems, the collected point cloud data inevitably contains some noise points, which affects the accuracy of the subsequent point cloud processing algorithm. However, the existing denoising algorithm will destroy the structural characteristics of the point cloud data while removing the noise, resulting in low accuracy of the final constructed model and the inability to obtain quality mapping data. Therefore, how to effectively retain edge features when denoising point cloud data to improve the accuracy of the three-dimensional terrain model has become a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a laser radar-based three-dimensional terrain mapping system and method, which can effectively retain edge features when denoising point cloud data to improve the accuracy of three-dimensional terrain models.

[0005] In a first aspect, the present application provides a three-dimensional terrain mapping method based on laser radar, the mapping method comprising the following steps:

[0006] Carry out terrain monitoring in the designated area and obtain the laser radar point cloud data of the designated area;

[0007] Determine the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data, construct an adaptive decision value according to the surveying and mapping features of all surveying and mapping points, perform edge screening on the surveying and mapping points in the laser radar point cloud data according to the adaptive decision value to obtain a surveying and mapping edge point set, and determine the surveying and mapping feature descriptor of each surveying and mapping point in the laser radar point cloud data through the surveying and mapping edge point set;

[0008] Perform incremental recovery on each surveying and mapping feature descriptor to obtain a coordinate increment corresponding to each surveying and mapping feature descriptor, reconstruct the surveying and mapping points in the laser radar point cloud data according to the corresponding coordinate increment, and thereby obtain reconstructed laser radar point cloud data of a specified area;

[0009] A three-dimensional terrain model of a designated area is constructed by the reconstructed laser radar point cloud data, and three-dimensional terrain mapping data of the designated area is obtained based on the three-dimensional terrain model.

[0010] In this embodiment, the terrain of a designated area is monitored by using a laser radar carried by a drone.

[0011] In this embodiment, determining the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data specifically includes:

[0012] For each surveying point in the laser radar point cloud data, selecting a neighboring point domain of the surveying point in the laser radar point cloud data according to a preset order of magnitude;

[0013] Performing singular value decomposition on the covariance matrix of the neighboring point domain, thereby obtaining a plurality of eigenvalues ​​corresponding to the surveying and mapping points;

[0014] The surveying and mapping features of the surveying and mapping points are determined according to all the eigenvalues, thereby obtaining the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data.

[0015] In this embodiment, constructing the adaptive decision value according to the surveying and mapping features of all surveying and mapping points specifically includes:

[0016] Determine the characteristic distribution center point and characteristic dispersion through the surveying and mapping characteristics of all surveying and mapping points;

[0017] An adaptive decision value is constructed using the feature distribution center point and the feature dispersion.

[0018] In this embodiment, edge screening is performed on the surveying points in the laser radar point cloud data according to the adaptive decision value to obtain a surveying edge point set, which specifically includes:

[0019] A set consisting of all surveying points in the laser radar point cloud data whose surveying characteristics are greater than the adaptive decision value is taken as a surveying edge point set.

[0020] In this embodiment, determining the surveying feature descriptor of each surveying point in the laser radar point cloud data by using the surveying edge point set specifically includes:

[0021] Determine a filtered normal vector of each surveying point in the laser radar point cloud data by using the surveying edge point set;

[0022] A mapping feature descriptor for each mapping point in the lidar point cloud data is constructed using the corresponding filtered normal vector.

[0023] In this embodiment, performing incremental recovery on each surveying and mapping feature descriptor to obtain the coordinate increment corresponding to each surveying and mapping feature descriptor specifically includes:

[0024] For each surveying and mapping feature descriptor, obtain the filter normal vector corresponding to the surveying and mapping feature descriptor;

[0025] The surveying and mapping feature descriptors are incrementally restored according to the filtered normal vector to obtain a coordinate increment corresponding to the surveying and mapping feature descriptors, and then the coordinate increment corresponding to each surveying and mapping feature descriptor is obtained.

[0026] In this embodiment, the surveying points in the laser radar point cloud data are reconstructed according to the corresponding coordinate increments to obtain the reconstructed laser radar point cloud data of the specified area, specifically including:

[0027] For each surveying point in the laser radar point cloud data, obtaining a coordinate increment corresponding to the surveying point;

[0028] The sum of the coordinate increment and the surveying point is used as the reconstructed surveying point, thereby obtaining all the reconstructed surveying points;

[0029] The data composed of all reconstructed surveying points is used as the reconstructed lidar point cloud data of the specified area.

[0030] In this embodiment, constructing a three-dimensional terrain model of a specified area using the reconstructed lidar point cloud data is to input the reconstructed lidar point cloud data into a point cloud visualization tool to generate a three-dimensional terrain model of the specified area.

[0031] In a second aspect, the present application provides a three-dimensional terrain surveying and mapping system based on laser radar, which is used to perform a three-dimensional terrain surveying and mapping method based on laser radar, and the surveying and mapping system includes:

[0032] The terrain monitoring module is used to monitor the terrain of a specified area and obtain the laser radar point cloud data of the specified area;

[0033] a descriptor extraction module, for determining the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data, constructing an adaptive decision value according to the surveying and mapping features of all surveying and mapping points, performing edge screening on the surveying and mapping points in the laser radar point cloud data according to the adaptive decision value, obtaining a surveying and mapping edge point set, and determining a surveying and mapping feature descriptor of each surveying and mapping point in the laser radar point cloud data through the surveying and mapping edge point set;

[0034] A data reconstruction module is used to incrementally restore each surveying and mapping feature descriptor to obtain a coordinate increment corresponding to each surveying and mapping feature descriptor, and reconstruct the surveying and mapping points in the laser radar point cloud data according to the corresponding coordinate increment, thereby obtaining the reconstructed laser radar point cloud data of the specified area;

[0035] The surveying and mapping data acquisition module is used to construct a three-dimensional terrain model of the specified area based on the reconstructed laser radar point cloud data, and obtain three-dimensional terrain surveying and mapping data of the specified area based on the three-dimensional terrain model.

[0036] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

[0037] The laser radar point cloud data of the designated area is obtained by performing terrain monitoring on the designated area; the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data are determined, an adaptive decision value is constructed according to the surveying and mapping features of all surveying and mapping points, and the edge of the surveying and mapping points in the laser radar point cloud data is screened according to the adaptive decision value to obtain a surveying and mapping edge point set, and the surveying and mapping feature descriptor of each surveying and mapping point in the laser radar point cloud data is determined by the surveying and mapping edge point set; each surveying and mapping feature descriptor is incrementally restored to obtain the coordinate increment corresponding to each surveying and mapping feature descriptor, and the surveying and mapping points in the laser radar point cloud data are reconstructed according to the corresponding coordinate increment to obtain the reconstructed laser radar point cloud data of the designated area; a three-dimensional terrain model of the designated area is constructed by the reconstructed laser radar point cloud data, and the three-dimensional terrain surveying and mapping data of the designated area is obtained based on the three-dimensional terrain model.

[0038] It can be seen that in this application, firstly, by filtering edge points through adaptive decision values, key features can be effectively retained during denoising, avoiding loss of details due to blind denoising, and the mapping feature descriptor combined with geometric attributes and edge point information can more comprehensively describe complex terrain and ensure high-precision expression of point clouds in local and overall features; then, the coordinate increment is calculated based on the mapping feature descriptor and the filter normal vector, which can distinguish random noise and greatly reduce outliers and random noise interference in the point cloud data, while avoiding loss of terrain details due to excessive smoothing, and using the filter normal vector to adjust the coordinate increment of the mapping point, so that the geometric relationship of the edge features of the point cloud can be accurately restored, and ultimately the accuracy, integrity and consistency of the model are effectively improved; finally, a more accurate and high-quality three-dimensional terrain model can be generated by reconstructing the lidar point cloud data, so that accurate three-dimensional terrain mapping data can be extracted from the three-dimensional terrain model.

[0039] In summary, the technical solution adopted in this application can effectively retain edge features when denoising point cloud data to improve the accuracy of the three-dimensional terrain model. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0041] Figure 1 is a flow chart of a three-dimensional terrain mapping method based on laser radar provided in this application;

[0042] Figure 2 is an exemplary flow chart for determining the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data provided by the present application;

[0043] Figure 3 is an exemplary flow chart for determining the coordinate increments corresponding to each mapping feature descriptor provided in the present application;

[0044] Figure 4 This is a module structure diagram of a three-dimensional terrain mapping system based on laser radar provided in this application. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0046] The embodiment of the present application provides a three-dimensional terrain mapping system and method based on laser radar, the core of which is to monitor the terrain of a designated area and obtain the laser radar point cloud data of the designated area; determine the mapping features of each mapping point in the laser radar point cloud data, build an adaptive decision value based on the mapping features of all mapping points, perform edge screening on the mapping points in the laser radar point cloud data according to the adaptive decision value, obtain a mapping edge point set, and determine the mapping feature descriptor of each mapping point in the laser radar point cloud data through the mapping edge point set; perform incremental recovery on each mapping feature descriptor to obtain the coordinate increment corresponding to each mapping feature descriptor, reconstruct the mapping points in the laser radar point cloud data according to the corresponding coordinate increment, and then obtain the reconstructed laser radar point cloud data of the designated area; construct a three-dimensional terrain model of the designated area from the reconstructed laser radar point cloud data, and obtain the three-dimensional terrain mapping data of the designated area based on the three-dimensional terrain model. The above scheme can effectively retain the edge features when denoising the point cloud data to improve the accuracy of the three-dimensional terrain model. Embodiment 1

[0047] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 As shown in FIG. 1 , this figure is an exemplary flow chart of a three-dimensional terrain mapping method based on laser radar according to this embodiment of the present application, and the mapping method includes the following steps:

[0048] In step S1, terrain monitoring is performed on a designated area to obtain lidar point cloud data of the designated area.

[0049] In specific implementation, the terrain of a designated area is monitored by a laser radar carried by a drone, thereby obtaining laser radar point cloud data of the designated area; the laser radar point cloud data contains a large number of surveying and mapping points in the designated area.

[0050] In step S2, the surveying and mapping features of each surveying and mapping point in the lidar point cloud data are determined, an adaptive decision value is constructed according to the surveying and mapping features of all surveying and mapping points, and the surveying and mapping points in the lidar point cloud data are edge screened according to the adaptive decision value to obtain a surveying and mapping edge point set, and the surveying and mapping feature descriptor of each surveying and mapping point in the lidar point cloud data is determined through the surveying and mapping edge point set.

[0051] Preferably, in this embodiment, reference Figure 2 As shown in FIG. 1 , this figure is an exemplary flow chart of determining the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data in an embodiment of the present application. In this embodiment, determining the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data can be implemented by the following steps:

[0052] First, in step S21, for each surveying point in the laser radar point cloud data, a neighboring point domain of the surveying point is selected in the laser radar point cloud data according to a preset order of magnitude;

[0053] Then, in step S22, singular value decomposition is performed on the covariance matrix of the neighboring point domain to obtain a plurality of eigenvalues ​​corresponding to the surveying point;

[0054] Finally, in step S23, the surveying and mapping features of the surveying and mapping points are determined according to all the characteristic values, thereby obtaining the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data.

[0055] In the specific implementation, first, for each surveying point in the laser radar point cloud data, a neighboring point domain of the surveying point is selected in the laser radar point cloud data according to a preset order of magnitude, wherein the neighboring point domain is a data domain containing the surveying point and its neighboring surveying points, and the order of magnitude is preset to 30 in the present application, that is, 30 neighboring surveying points around the surveying point are selected, which is not limited here; then, a covariance matrix of the neighboring point domain can be constructed, and then a singular value decomposition is performed on the covariance matrix of the neighboring point domain, and multiple eigenvalues ​​corresponding to the surveying point can be obtained; finally, the surveying feature of the surveying point is determined according to all the eigenvalues, the smallest eigenvalue can be selected, and the sum of all the eigenvalues ​​can be calculated, so that the ratio of the smallest eigenvalue to the calculated result can be used as the surveying feature of the surveying point, and the surveying feature represents the degree of curvature of the point cloud surface at the position of the surveying point. The surveying feature of each surveying point in the laser radar point cloud data can be obtained in the above manner.

[0056] In this embodiment, the adaptive decision value can be constructed according to the surveying and mapping features of all surveying and mapping points in the following manner, namely:

[0057] Determine the characteristic distribution center point and characteristic dispersion through the surveying and mapping characteristics of all surveying and mapping points;

[0058] An adaptive decision value is constructed using the feature distribution center point and the feature dispersion.

[0059] In the specific implementation, first, the feature distribution center point and feature discreteness are determined through the surveying and mapping features of all surveying and mapping points, wherein the feature distribution center point refers to the center point of all surveying and mapping features under the overall distribution, and the feature discreteness represents the degree of discreteness of all surveying and mapping features. The mean of all surveying and mapping features can be used as the value of the feature distribution center point, and the standard deviation of all surveying and mapping features can be used as the feature discreteness; then, the feature distribution center point and feature discreteness can be used to construct an adaptive decision value, which is a decision value for screening edge surveying and mapping points, and the sum of the feature distribution center point and the feature discreteness can be used as the adaptive decision value.

[0060] In this embodiment, the edge screening of the surveying points in the laser radar point cloud data is performed according to the adaptive decision value to obtain the surveying edge point set, which can be specifically carried out in the following manner, namely:

[0061] A set consisting of all surveying points in the laser radar point cloud data whose surveying characteristics are greater than the adaptive decision value is taken as a surveying edge point set.

[0062] It should be noted that usually the surveying points with larger surveying features are located at the edge or outlier positions, and these positions are prone to normal vector estimation errors.

[0063] In this embodiment, the mapping feature descriptor of each mapping point in the laser radar point cloud data is determined by the mapping edge point set in the following manner, namely:

[0064] Determine a filtered normal vector of each surveying point in the laser radar point cloud data by using the surveying edge point set;

[0065] A mapping feature descriptor for each mapping point in the lidar point cloud data is constructed using the corresponding filtered normal vector.

[0066] In specific implementation, the filtered normal vector of each surveying point in the laser radar point cloud data can be determined by the surveying edge point set, and all surveying points in the surveying edge point set are sorted in the order of ascending surveying features. Smaller surveying features represent that the surveying point is on a relatively flat surface. For each surveying point a in the surveying edge point set, the surveying point closest to a is found in the surveying edge point set and the non-edge point set respectively, and the dot product of the normal vectors corresponding to the two selected surveying points and a is calculated, and the normal vector of a is replaced with the normal vector direction corresponding to the point with a smaller dot product value, and the surveying point a is deleted from the surveying edge point set. The above process will be iteratively performed in the order of ascending surveying features in the surveying edge point set until the normal vectors of all points in the surveying edge point set are updated and the final filtered normal vector is obtained, that is, the filtered normal vector of each surveying point in the laser radar point cloud data is obtained. After the normal vector direction is adjusted by the adaptive decision value method, the normal vector at the outlier position is consistent with the normal vector direction of the surveying point on the adjacent plane, which can solve the problem of incorrect normal vector estimation of a few surveying points.

[0067] In addition, in a specific implementation, the corresponding filtered normal vector can be used to construct a surveying feature descriptor for each surveying point in the laser radar point cloud data, wherein the surveying feature descriptor represents a descriptor of the projection distribution feature of the surveying point in the direction of the corresponding normal vector; in actual implementation, for each surveying point in the laser radar point cloud data, the tangent plane of the surveying point can be calculated by the filtered normal vector of the surveying point, and the tangent plane is perpendicular to the filtered normal vector. The projection coordinates of each neighboring surveying point on the tangent plane and the projection value on the filtered normal vector can be calculated by using the tangent plane and the filtered normal vector of the surveying point position. In actual implementation, the projection coordinates of each neighboring surveying point on the tangent plane and the projection value on the filtered normal vector can be calculated by the following method, namely:

[0068]

[0069] in, represents the i-th survey point, Indicates the survey point The jth nearest neighbor mapping point, Indicates neighboring survey points The projection coordinates on the tangent plane, Indicates the survey point The filtered normal vector, Indicates neighboring survey points The projection value on the filtered normal vector is used to calculate the distance from the surveying point based on the projection coordinates of each neighboring surveying point on the tangent plane. The nearest n neighboring surveying points are selected, and the value of n can be determined through historical experiments. The projection values ​​of these n neighboring surveying points on the filtered normal vector are arranged according to the distance, so that the arrangement result can be used as the surveying feature descriptor of the surveying point. The surveying feature descriptor of each surveying point in the lidar point cloud data can be obtained in the above manner.

[0070] It should be noted that by filtering edge points through adaptive decision values, key features can be effectively retained during denoising, avoiding the loss of details due to blind denoising. The surveying and mapping feature descriptor combines geometric attributes and edge point information to more comprehensively describe complex terrain and ensure high-precision expression of point clouds in local and overall features.

[0071] In step S3, each surveying and mapping feature descriptor is incrementally restored to obtain the coordinate increment corresponding to each surveying and mapping feature descriptor, and the surveying and mapping points in the lidar point cloud data are reconstructed according to the corresponding coordinate increment to obtain the reconstructed lidar point cloud data of the specified area.

[0072] Preferably, in this embodiment, reference Figure 3 As shown, this figure is an exemplary flow chart of determining the coordinate increments corresponding to each surveying and mapping feature descriptor in an embodiment of the present application. In this embodiment, each surveying and mapping feature descriptor is incrementally restored to obtain the coordinate increments corresponding to each surveying and mapping feature descriptor, which can be specifically implemented by the following steps:

[0073] First, in step S31, for each surveying and mapping feature descriptor, a filter normal vector corresponding to the surveying and mapping feature descriptor is obtained;

[0074] Then, in step S32, the surveying and mapping feature descriptor is incrementally restored according to the filtered normal vector to obtain the coordinate increment corresponding to the surveying and mapping feature descriptor, and then the coordinate increment corresponding to each surveying and mapping feature descriptor is obtained.

[0075] In specific implementation, first, for each mapping feature descriptor, the filter normal vector corresponding to the mapping feature descriptor can be obtained by traversal, that is, the filter normal vector of the mapping point corresponding to the mapping feature descriptor can be used as the filter normal vector corresponding to the mapping feature descriptor; then, the mapping feature descriptor is incrementally restored according to the filter normal vector to obtain the coordinate increment corresponding to the mapping feature descriptor, wherein the coordinate increment represents the coordinate translation amount required for the mapping point in the direction of the corresponding filter normal vector. In actual implementation, the coordinate increment can be determined by the following method:

[0076]

[0077] in, represents the coordinate increment corresponding to the i-th mapping feature descriptor, represents the filter normal vector corresponding to the i-th mapping feature descriptor, represents the projection value at the vth position in the i-th mapping feature descriptor, n represents the total number of projection values ​​in, Represents the bilateral Gaussian weight, which can be set through historical experiments. The coordinate increment corresponding to each mapping feature descriptor can be obtained through the above method.

[0078] In this embodiment, the surveying points in the laser radar point cloud data are reconstructed according to the corresponding coordinate increments, and the reconstructed laser radar point cloud data of the specified area can be obtained in the following manner, namely:

[0079] For each surveying point in the laser radar point cloud data, obtaining a coordinate increment corresponding to the surveying point;

[0080] The sum of the coordinate increment and the surveying point is used as the reconstructed surveying point, thereby obtaining all the reconstructed surveying points;

[0081] The data composed of all reconstructed surveying points is used as the reconstructed lidar point cloud data of the specified area.

[0082] In the specific implementation, first, for each surveying point in the lidar point cloud data, the coordinate increment corresponding to the surveying point can be obtained, that is, the coordinate increment of the surveying feature descriptor corresponding to the surveying point can be used as the coordinate increment corresponding to the surveying point; then, the sum of the coordinate increment and the surveying point can be used as the reconstructed surveying point, and the reconstructed surveying point is the reconstructed surveying point with noise removed, thereby obtaining all the reconstructed surveying points; finally, the data composed of all the reconstructed surveying points can be used as the reconstructed lidar point cloud data of the specified area.

[0083] It should be noted that the coordinate increment is calculated based on the surveying and mapping feature descriptors and filtered normal vectors, which can distinguish random noise and make targeted adjustments, greatly reducing the interference of outliers and random noise in the point cloud data, while avoiding the loss of terrain details due to excessive smoothing, and using filtered normal vectors to adjust the coordinate increments of the surveying points, so that the geometric relationship of the point cloud edge features can be accurately restored, ultimately effectively improving the accuracy, completeness and consistency of the model.

[0084] In step S4, a three-dimensional terrain model of the designated area is constructed based on the reconstructed lidar point cloud data, and three-dimensional terrain mapping data of the designated area is obtained based on the three-dimensional terrain model.

[0085] In this embodiment, constructing a three-dimensional terrain model of a specified area using the reconstructed lidar point cloud data is to input the reconstructed lidar point cloud data into a point cloud visualization tool to generate a three-dimensional terrain model of the specified area.

[0086] In this embodiment, three-dimensional terrain mapping data of a specified area is obtained based on the three-dimensional terrain model; it should be noted that, in the present application, the three-dimensional terrain mapping data includes a three-dimensional terrain model file, a contour map, a profile map and an elevation distribution map of the specified area; in specific implementation, contour data can be extracted from the three-dimensional terrain model to generate a two-dimensional outline representation of the terrain, i.e., a contour map; an elevation distribution map can be generated according to the Z coordinate of the three-dimensional terrain model to intuitively display the terrain undulating features; and profile data can be extracted from the three-dimensional terrain model according to a specified path or area to generate a terrain profile map.

[0087] It should be noted that by reconstructing the lidar point cloud data, a more accurate and high-quality three-dimensional terrain model can be generated, so that accurate three-dimensional terrain mapping data can be extracted from the three-dimensional terrain model.

[0088] It can be seen that in this application, firstly, by filtering edge points through adaptive decision values, key features can be effectively retained during denoising, avoiding loss of details due to blind denoising, and the mapping feature descriptor combined with geometric attributes and edge point information can more comprehensively describe complex terrain and ensure high-precision expression of point clouds in local and overall features; then, the coordinate increment is calculated based on the mapping feature descriptor and the filter normal vector, which can distinguish random noise and greatly reduce outliers and random noise interference in the point cloud data, while avoiding loss of terrain details due to excessive smoothing, and using the filter normal vector to adjust the coordinate increment of the mapping point, so that the geometric relationship of the edge features of the point cloud can be accurately restored, and ultimately the accuracy, integrity and consistency of the model are effectively improved; finally, a more accurate and high-quality three-dimensional terrain model can be generated by reconstructing the lidar point cloud data, so that accurate three-dimensional terrain mapping data can be extracted from the three-dimensional terrain model.

[0089] In summary, the technical solution adopted in this application can effectively retain edge features when denoising point cloud data to improve the accuracy of the three-dimensional terrain model.

[0090] Embodiment 2

[0091] This application provides a three-dimensional terrain mapping system based on laser radar. Figure 4 As shown, this figure is a schematic diagram of a surveying and mapping system according to this embodiment of the present application, and the surveying and mapping system includes:

[0092] The terrain monitoring module 100 is used to monitor the terrain of a designated area and obtain the laser radar point cloud data of the designated area;

[0093] A descriptor extraction module 200 is used to determine the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data, construct an adaptive decision value according to the surveying and mapping features of all surveying and mapping points, perform edge screening on the surveying and mapping points in the laser radar point cloud data according to the adaptive decision value, obtain a surveying and mapping edge point set, and determine the surveying and mapping feature descriptor of each surveying and mapping point in the laser radar point cloud data through the surveying and mapping edge point set;

[0094] The data reconstruction module 300 is used to incrementally restore each surveying and mapping feature descriptor to obtain the coordinate increment corresponding to each surveying and mapping feature descriptor, and reconstruct the surveying and mapping points in the laser radar point cloud data according to the corresponding coordinate increment, thereby obtaining the reconstructed laser radar point cloud data of the specified area;

[0095] The surveying and mapping data acquisition module 400 is used to construct a three-dimensional terrain model of a specified area based on the reconstructed lidar point cloud data, and obtain three-dimensional terrain surveying and mapping data of the specified area based on the three-dimensional terrain model.

[0096] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0097] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, the storage medium including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically-erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0098] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A three-dimensional terrain mapping method based on laser radar, characterized in that: The surveying and mapping method comprises the following steps: Carry out terrain monitoring in the designated area and obtain the laser radar point cloud data of the designated area; Determine the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data, construct an adaptive decision value according to the surveying and mapping features of all surveying and mapping points, perform edge screening on the surveying and mapping points in the laser radar point cloud data according to the adaptive decision value to obtain a surveying and mapping edge point set, and determine the surveying and mapping feature descriptor of each surveying and mapping point in the laser radar point cloud data through the surveying and mapping edge point set; Perform incremental recovery on each surveying and mapping feature descriptor to obtain a coordinate increment corresponding to each surveying and mapping feature descriptor, reconstruct the surveying and mapping points in the laser radar point cloud data according to the corresponding coordinate increment, and thereby obtain reconstructed laser radar point cloud data of a specified area; Constructing a three-dimensional terrain model of a designated area from the reconstructed laser radar point cloud data, and obtaining three-dimensional terrain surveying and mapping data of the designated area based on the three-dimensional terrain model; Wherein, determining the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data specifically includes: For each surveying point in the laser radar point cloud data, selecting a neighboring point domain of the surveying point in the laser radar point cloud data according to a preset order of magnitude; Performing singular value decomposition on the covariance matrix of the neighboring point domain, thereby obtaining a plurality of eigenvalues ​​corresponding to the surveying and mapping points; Determine the surveying and mapping features of the surveying and mapping points according to all the eigenvalues, and then obtain the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data; Among them, constructing adaptive decision values ​​according to the surveying and mapping characteristics of all surveying and mapping points specifically includes: Determine the characteristic distribution center point and characteristic dispersion through the surveying and mapping characteristics of all surveying and mapping points; Constructing an adaptive decision value using the feature distribution center point and the feature dispersion; Wherein, determining the surveying feature descriptor of each surveying point in the laser radar point cloud data through the surveying edge point set specifically includes: Determine a filtered normal vector of each surveying point in the laser radar point cloud data by using the surveying edge point set; A mapping feature descriptor for each mapping point in the lidar point cloud data is constructed using the corresponding filtered normal vector.

2. A three-dimensional terrain mapping method based on laser radar as claimed in claim 1, characterized in that: The terrain of the designated area is monitored through the laser radar carried by the drone.

3. The three-dimensional terrain mapping method based on laser radar according to claim 1, characterized in that: Performing edge screening on the surveying points in the laser radar point cloud data according to the adaptive decision value to obtain a surveying edge point set specifically includes: A set consisting of all surveying points in the laser radar point cloud data whose surveying characteristics are greater than the adaptive decision value is taken as a surveying edge point set.

4. The three-dimensional terrain mapping method based on laser radar according to claim 1, characterized in that: Incremental recovery is performed on each surveying and mapping feature descriptor to obtain the coordinate increments corresponding to each surveying and mapping feature descriptor, including: For each surveying and mapping feature descriptor, obtain the filter normal vector corresponding to the surveying and mapping feature descriptor; The surveying and mapping feature descriptors are incrementally restored according to the filtered normal vector to obtain a coordinate increment corresponding to the surveying and mapping feature descriptors, and then the coordinate increment corresponding to each surveying and mapping feature descriptor is obtained.

5. The three-dimensional terrain mapping method based on laser radar according to claim 1, characterized in that: Reconstructing the surveying points in the laser radar point cloud data according to the corresponding coordinate increments, and then obtaining the reconstructed laser radar point cloud data of the specified area specifically includes: For each surveying point in the laser radar point cloud data, obtaining a coordinate increment corresponding to the surveying point; The sum of the coordinate increment and the surveying point is used as the reconstructed surveying point, thereby obtaining all the reconstructed surveying points; The data composed of all reconstructed surveying points is used as the reconstructed lidar point cloud data of the specified area.

6. The three-dimensional terrain mapping method based on laser radar according to claim 1, characterized in that: Constructing a three-dimensional terrain model of a designated area from the reconstructed laser radar point cloud data is to input the reconstructed laser radar point cloud data into a point cloud visualization tool to generate a three-dimensional terrain model of the designated area.

7. A three-dimensional terrain mapping system based on laser radar, used to execute a three-dimensional terrain mapping method based on laser radar as claimed in any one of claims 1 to 6, characterized in that: The surveying and mapping system comprises: The terrain monitoring module is used to monitor the terrain of a specified area and obtain the laser radar point cloud data of the specified area; a descriptor extraction module, for determining the surveying and mapping features of each surveying and mapping point in the laser radar point cloud data, constructing an adaptive decision value according to the surveying and mapping features of all surveying and mapping points, performing edge screening on the surveying and mapping points in the laser radar point cloud data according to the adaptive decision value, obtaining a surveying and mapping edge point set, and determining a surveying and mapping feature descriptor of each surveying and mapping point in the laser radar point cloud data through the surveying and mapping edge point set; A data reconstruction module is used to incrementally restore each surveying and mapping feature descriptor to obtain a coordinate increment corresponding to each surveying and mapping feature descriptor, and reconstruct the surveying and mapping points in the laser radar point cloud data according to the corresponding coordinate increment, thereby obtaining the reconstructed laser radar point cloud data of the specified area; The surveying and mapping data acquisition module is used to construct a three-dimensional terrain model of the specified area based on the reconstructed laser radar point cloud data, and obtain three-dimensional terrain surveying and mapping data of the specified area based on the three-dimensional terrain model.

Images