An airborne lidar point cloud single tree segmentation method, device, equipment and medium

Abstract

The present application discloses a method, device, equipment and medium for single-tree segmentation of airborne lidar point clouds, which relates to the field of single-tree segmentation. The method includes: performing point cloud voxelization on the target point cloud to obtain cuboid voxels; extracting the modulus points in each cuboid voxel; generating a connectivity graph according to the modulus points in each cuboid voxel and the corresponding group of neighboring modulus points; extracting the initial tree base points from the connectivity graph; determining the material transportation paths of the modulus points in each cuboid voxel according to the connectivity graph and the initial tree base points; for any initial tree base point, determining the number of paths of the material transportation path passing through the initial tree base point; screening the initial tree base points according to the number of paths to obtain the final tree base points, and determining the single-tree segmentation result of the target forest according to the final tree base points and the material transportation paths passing through the final tree base points. The present application can improve the accuracy and applicability of single-tree segmentation.

Description

Technical Field

[0001] The present application relates to the field of individual tree segmentation, and in particular, to a method, device, equipment and medium for segmenting individual trees from airborne lidar point clouds. Background Art

[0002] Forests are crucial for maintaining ecological balance and sustainable development. Forest research at the individual tree scale helps to reveal the complexity and diversity of ecosystems and provides important support for forest management and ecological research. Airborne lidar technology accurately depicts the three-dimensional structure of forests with high-resolution point clouds and is a commonly used tool for obtaining information at the individual tree scale in large areas.

[0003] Accurately segmenting individual trees from lidar point clouds is the key to quantifying tree structures. Related individual tree segmentation methods are mainly divided into two categories: top-down methods and bottom-up methods.

[0004] Top-down methods analyze trees based on canopy characteristics, including the canopy height feature method and the density feature method. The canopy height feature method identifies the tree tops through a local maximum filter and combines region growing or watershed algorithms to segment the canopy. Such methods are suitable for coniferous forests with obvious canopy height characteristics. The canopy density feature method uses clustering algorithms to segment individual trees, and typical methods include Meanshift, K-means, and spectral clustering. Such methods are suitable for broad-leaved forests and understory vegetation segmentation, but are limited by the significance of the point cloud distribution characteristics and have weak segmentation stability.

[0005] Bottom-up methods detect individual trees by detecting the positions of tree trunks. Trunk recognition relies on point cloud density, intensity, shape, and vertical morphological characteristics. Compared with canopy characteristics, the trunk morphology has higher similarity, and the bottom-up method is more robust to tree species and leaf conditions.

[0006] Although the bottom-up method has better scene generality, it still faces the following challenges: the point cloud density method cannot eliminate the interference of understory vegetation; the intensity method is vulnerable to the influence of canopy occlusion on point cloud intensity in trunk recognition and shows unstable performance; the shape method is only applicable to high-density and high-precision point clouds that can record the cross-sectional shape of tree trunks; the vertical morphology method has insufficient applicability in mixed forests. Therefore, how to improve the accuracy and applicability of individual tree segmentation has become an urgent problem to be solved at present. Summary of the Invention

[0007] The purpose of the present application is to provide a method, device, equipment and medium for segmenting individual trees from airborne lidar point clouds to improve the accuracy and applicability of individual tree segmentation.

[0008] To achieve the above purpose, the present application provides the following solutions:

[0009] In a first aspect, the present application provides a method for segmenting individual trees from airborne lidar point clouds, including:

[0010] Obtain a target point cloud; the target point cloud is obtained by scanning a target forest using an airborne lidar.

[0011] Perform voxelization on the target point cloud to obtain cuboid voxels.

[0012] Extract the model points within each cuboid voxel; the model point is the laser point closest to the centroid of the cuboid voxel.

[0013] Generate a connectivity graph based on the model points within each cuboid voxel and the corresponding set of neighboring model points; the set of neighboring model points includes: the model points within the neighboring cuboid voxels of the cuboid voxel where the model point is located.

[0014] Extract initial tree base points from the connectivity graph; the initial tree base points include: the model points in the low-level voxels; the low-level voxels are the cuboid voxels at a set level in the connectivity graph.

[0015] Determine the material transportation path for each model point within each cuboid voxel based on the connectivity graph and the initial tree base points; the material transportation path is the shortest path in the path set of this model point; the path set includes: the shortest paths between this model point and each of the initial tree base points.

[0016] For any one of the initial tree base points, determine the number of paths of the material transportation path passing through this initial tree base point.

[0017] Screen the initial tree base points according to the number of paths to obtain the final tree base points, and determine the single-tree segmentation result of the target forest based on the final tree base points and the material transportation paths passing through the final tree base points.

[0018] Optionally, obtaining the target point cloud specifically includes:

[0019] Collect the initial lidar point cloud of the target forest using an airborne lidar.

[0020] Extract the ground points from the initial lidar point cloud using a cloth simulation filtering algorithm, and construct a terrain model based on the ground points.

[0021] Calculate the height difference between the initial lidar point cloud and the terrain model.

[0022] Use the height difference as the point cloud elevation, perform normalization processing on the height information of the initial lidar point cloud to obtain a height-normalized point cloud, and use the height-normalized point cloud as the target point cloud.

[0023] Optionally, performing voxelization on the target point cloud to obtain cuboid voxels specifically includes:

[0024] Voxelize the point cloud according to the three-dimensional coordinates of each laser point in the target point cloud to obtain cuboid voxels; the expression of the cuboid voxel is:

[0025]

[0026] where (r, c, l) represents the row-column-layer number of the cuboid voxel to which the laser point belongs, r represents the row number, c represents the column number, and l represents the layer number; int represents the integer function; (x, y, z) represents the three-dimensional coordinates of the laser point, x represents the x-axis coordinate value, y represents the y-axis coordinate value, and z represents the z-axis coordinate value; x min represents the minimum x-axis coordinate value of the target point cloud, y min represents the minimum y-axis coordinate value of the target point cloud, z min represents the minimum z-axis coordinate value of the target point cloud; v xy represents the horizontal dimension of the cuboid voxel; v z represents the vertical dimension of the cuboid voxel.

[0027] Optionally, extract the model points in each of the cuboid voxels, specifically including:

[0028] For any one of the cuboid voxels, calculate the Euclidean distance between each laser point in the cuboid voxel and the centroid of the cuboid voxel, and determine the laser point with the minimum Euclidean distance as the model point in the cuboid voxel.

[0029] Optionally, generate a connectivity graph according to the model points in each of the cuboid voxels and the corresponding group of neighboring model points, specifically including:

[0030] For the model point in any one of the cuboid voxels, determine the model points in the 26-neighboring cuboid voxels of the cuboid voxel where the model point is located as the group of neighboring model points of the model point;

[0031] For the model point in any one of the cuboid voxels, connect the model point and each of the model points in the corresponding group of neighboring model points to generate a topological graph of the model point;

[0032] Determine the connectivity graph according to the topological graphs of the model points in all the cuboid voxels.

[0033] Optionally, determine the material transportation path of the model point in each of the cuboid voxels according to the connectivity graph and the initial tree base point, specifically including:

[0034] For the model point in any one of the cuboid voxels, calculate the shortest path between the model point and each of the initial tree base points by using the Dijkstra algorithm according to the connectivity graph to obtain a path set of the model point;

[0035] For any modulus point within any of the cuboid voxels, select the shortest path from the set of paths to obtain the material transportation path of this modulus point.

[0036] Optionally, screen the initial tree base points according to the number of paths to obtain the final tree base points, and determine the single-tree segmentation result of the target forest according to the final tree base points and the material transportation paths passing through the final tree base points, specifically including:

[0037] Determine the initial tree base points with the number of paths exceeding the set threshold as the final tree base points;

[0038] For any of the final tree base points, determine all the material transportation paths passing through this final tree base point as the target path set of this final tree base point;

[0039] For any of the final tree base points, determine the cuboid voxel to which the modulus points corresponding to all the material transportation paths in the target path set belong as the target voxel corresponding to this final tree base point;

[0040] For any of the final tree base points, generate the single-tree segmentation result of this final tree base point according to the point cloud in the target voxel; the single-tree segmentation results of all the final tree base points are used as the single-tree segmentation result of the target forest.

[0041] In a second aspect, the present application provides an airborne lidar point cloud single-tree segmentation device, including:

[0042] A point cloud acquisition module, configured to acquire a target point cloud; the target point cloud is obtained by scanning a target forest using an airborne lidar;

[0043] A point cloud voxelization module, configured to perform point cloud voxelization on the target point cloud to obtain cuboid voxels;

[0044] A modulus point extraction module, configured to extract modulus points within each of the cuboid voxels; the modulus point is the laser point closest to the centroid of the cuboid voxel;

[0045] A connected graph generation module, configured to generate a connected graph according to the modulus points within each of the cuboid voxels and the corresponding set of neighboring modulus points; the set of neighboring modulus points includes: the modulus points within the neighboring cuboid voxels of the cuboid voxel where this modulus point is located;

[0046] An initial tree base point extraction module, configured to extract initial tree base points from the connected graph; the initial tree base points include: the modulus points in the low-level voxels; the low-level voxels are the cuboid voxels at the set layer in the connected graph;

[0047] A material transportation path determination module, configured to determine the material transportation path of the module points in each of the cuboid voxels according to the connectivity graph and the initial tree base points; the material transportation path is the shortest path in the path set of the module point; the path set includes: the shortest paths between the module point and each of the initial tree base points;

[0048] A path quantity determination module, configured to determine, for any one of the initial tree base points, the quantity of the material transportation paths passing through the initial tree base point;

[0049] A single-tree segmentation result determination module, configured to screen the initial tree base points according to the path quantity to obtain the final tree base points, and determine the single-tree segmentation result of the target forest according to the final tree base points and the material transportation paths passing through the final tree base points.

[0050] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the airborne lidar point cloud single-tree segmentation method described in any one of the above.

[0051] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the airborne lidar point cloud single-tree segmentation method described in any one of the above.

[0052] According to the specific embodiments provided by the present application, the present application has the following technical effects:

[0053] The present application provides an airborne lidar point cloud single-tree segmentation method, device, equipment and medium. By performing point cloud voxelization on the lidar point cloud to obtain cuboid voxels, extracting the laser points closest to the centroid in each cuboid voxel as module points, and generating a connectivity graph according to the module points and the corresponding adjacent module point groups in each cuboid voxel, the constructed connectivity graph has stronger vertical connectivity. When it is used for single-tree segmentation, it can significantly improve the accuracy of single-tree segmentation; by extracting the initial tree base points from the connectivity graph, determining the material transportation path of the module points in each cuboid voxel according to the connectivity graph and the initial tree base points, and performing single-tree segmentation based on the material transportation path, it is applicable to a variety of complex forest scenarios, ensuring high applicability of single-tree segmentation under different conditions. <{ Description of the Drawings

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to these drawings.

[0055] Figure 1 It is an application environment diagram of a method for segmenting individual trees from airborne lidar point clouds in an embodiment of the present application;

[0056] Figure 2 It is a schematic flowchart of a method for segmenting individual trees from airborne lidar point clouds provided in an embodiment of the present application;

[0057] Figure 3 It is an effect diagram of key steps provided in an embodiment of the present application;

[0058] Figure 4 It is a schematic diagram of the individual tree segmentation results in different forest scenarios provided in an embodiment of the present application;

[0059] Figure 5 It is a schematic diagram of the functional modules of a device for segmenting individual trees from airborne lidar point clouds provided in another embodiment of the present application;

[0060] Figure 6 It is a schematic diagram of the structure of a computer device provided in an embodiment of the present application. Detailed implementation manners

[0061] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0062] To make the above objects, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below in conjunction with the drawings and specific implementation manners.

[0063] The method for segmenting individual trees from airborne lidar point clouds provided in the embodiments of the present application can be applied to, for example Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set separately, integrated on the server 104, placed on the cloud or other servers. The terminal 102 can send the target point cloud to the server 104. After receiving the target point cloud, for the target point cloud, the server 104 performs point cloud voxelization on the target point cloud to obtain cuboid voxels; extracts the modulus points within each cuboid voxel; generates a connectivity graph according to the modulus points within each cuboid voxel and the corresponding adjacent modulus point groups; extracts the initial tree base points from the connectivity graph; determines the material transportation paths of the modulus points within each cuboid voxel according to the connectivity graph and the initial tree base points; for any initial tree base point, determines the number of paths of the material transportation path passing through the initial tree base point; screens the initial tree base points according to the number of paths to obtain the final tree base points, and determines the single-tree segmentation result of the target forest according to the final tree base points and the material transportation paths passing through the final tree base points.

[0064] The server 104 can feedback the obtained single-tree segmentation result of the target forest to the terminal 102. In addition, in some embodiments, the airborne lidar point cloud single-tree segmentation method can also be implemented separately by the server 104 or the terminal 102. For example, the terminal 102 can directly process the target point cloud, or the server 104 can obtain the target point cloud from the data storage system and process the target point cloud.

[0065] Among them, the terminal 102 can be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers, and can also be a cloud server.

[0066] In an exemplary embodiment, as Figure 2 shown, an airborne lidar point cloud single-tree segmentation method is provided. This method is executed by a computer device, and can be specifically executed alone by a computer device such as a terminal or a server, or jointly executed by a terminal and a server. In the embodiments of the present application, taking this method applied to Figure 1 the server 104 in

[0067] Step 201, obtain a target point cloud.

[0068] Among them, the target point cloud is obtained by scanning a target forest using an airborne lidar, and the target point cloud includes a plurality of laser points.

[0069] Step 202: Voxelize the target point cloud to obtain cuboid voxels.

[0070] Step 203: Extract the model points within each of the cuboid voxels.

[0071] Wherein, the model point is the laser point closest to the centroid of the cuboid voxel.

[0072] Step 204: Generate a connectivity graph based on the model points within each cuboid voxel and the corresponding neighboring model point groups.

[0073] Wherein, the neighboring model point group includes: the model points within the neighboring cuboid voxels of the cuboid voxel where the model point is located.

[0074] Step 205: Extract the initial tree base points from the connectivity graph.

[0075] The initial tree base points include: the model points in the low-level voxels; the low-level voxels are the cuboid voxels at a set layer number in the connectivity graph.

[0076] Step 206: Determine the material transportation path of the model points within each cuboid voxel according to the connectivity graph and the initial tree base points.

[0077] The material transportation path is the shortest path in the path set of the model point; the path set includes: the shortest paths between the model point and each of the initial tree base points.

[0078] Step 207: For any one of the initial tree base points, determine the number of paths of the material transportation path passing through the initial tree base point.

[0079] The purpose of determining the number of paths is to evaluate the cumulative intensity of each initial tree base point for subsequent screening of the tree base points.

[0080] Step 208: Screen the initial tree base points according to the number of paths to obtain the final tree base points, and determine the single-tree segmentation result of the target forest according to the final tree base points and the material transportation paths passing through the final tree base points.

[0081] Implementing the above steps 201 to 208 can improve the accuracy and applicability of single-tree segmentation.

[0082] In another exemplary embodiment of the present application, step 201 specifically includes:

[0083] Collect the initial lidar point cloud of the target forest using airborne lidar; extract the ground points from the initial lidar point cloud using a cloth simulation filtering algorithm, and construct a terrain model based on the ground points; calculate the height difference between the initial lidar point cloud and the terrain model; use the height difference as the point cloud elevation to normalize the height information of the initial lidar point cloud to obtain a height-normalized point cloud, and use the height-normalized point cloud as the target point cloud. Among them, the process of normalization is: replace the height information of the initial lidar point cloud with the height difference.

[0084] In another exemplary embodiment of the present application, step 202 specifically includes:

[0085] Voxelize the point cloud according to the three-dimensional coordinates of each laser point in the target point cloud to obtain cuboid voxels, forming the basic unit for mold point extraction. The expression of the cuboid voxel is:

[0086]

[0087] Among them, (r, c, l) represents the row-column-layer number of the cuboid voxel to which the laser point belongs, r represents the row number, c represents the column number, and l represents the layer number; int represents the integer function; (x, y, z) represents the three-dimensional coordinates of the laser point, x represents the x-axis coordinate value, y represents the y-axis coordinate value, and z represents the z-axis coordinate value; x min , y min and z min represent the minimum x, y, z values of the target point cloud, that is, x min represents the minimum x-axis coordinate value of the target point cloud, y min represents the minimum y-axis coordinate value of the target point cloud, z min represents the minimum z-axis coordinate value of the target point cloud; v xy represents the horizontal dimension of the cuboid voxel; v z represents the vertical dimension of the cuboid voxel.

[0088] In another exemplary embodiment of the present application, step 203 specifically includes:

[0089] For any one of the cuboid voxels, calculate the Euclidean distance between each laser point in the cuboid voxel and the centroid of the cuboid voxel, and determine the laser point with the minimum Euclidean distance as the mold point in the cuboid voxel, thus providing a basis for the construction of the connected graph.

[0090] In each cuboid voxel of this embodiment, the laser point closest to the centroid is selected as the mold point, and its calculation formula is as follows:

[0091] argmin||p i -c|| (2)

[0092] where argmin represents the index function of the modulus point v, ||·|| represents the Euclidean distance, p i is the coordinate of the i-th laser point, and c represents the coordinate of the centroid point. All modulus points are stored in the modulus point set V, that is, v ∈ V.

[0093] In another exemplary embodiment of the present application, step 204 specifically includes:

[0094] For any modulus point within any of the cuboid voxels, the modulus points within the 26-neighbor cuboid voxels of the cuboid voxel where the modulus point is located are determined as the neighboring modulus point group of the modulus point; for any modulus point within any of the cuboid voxels, the modulus point and each modulus point in the corresponding neighboring modulus point group are connected to each other to generate a topological graph of the modulus point; a connected graph is determined according to the topological graphs of all the modulus points within the cuboid voxels. The connected graph is used to describe the topological relationship between point clouds.

[0095] In this embodiment, the calculation formula for any neighboring modulus point v' in the neighboring modulus point group is as follows:

[0096] {(r', c', l') || |r - r'| ≤ 1, |c - c'| ≤ 1, |l - l'| ≤ 1, (r', c', l') ≠ (r, c, l)} (3)

[0097] where (r, c, l) and (r', c', l') respectively represent the row, column, and layer numbers of v and v'. The weight of the connecting edge e between v and v' is the Euclidean distance ||v - v'|| between the two points. All connecting edges are stored in the edge set E, that is, e ∈ E.

[0098] In another exemplary embodiment of the present application, in step 205, the modulus points of the lower-layer voxels are extracted from the connected graph as the initial tree base points to calibrate the potential trunk starting points. The lower-layer voxels are the cuboid voxels in the second layer from bottom to top in the connected graph.

[0099] In another exemplary embodiment of the present application, step 206 specifically includes:

[0100] (1) For any modulus point within any of the cuboid voxels, the shortest path between the modulus point and each of the initial tree base points is calculated according to the connected graph by using the Dijkstra algorithm, and a path set of the modulus point is obtained.

[0101] The shortest path between the modulus point and the initial tree base point is determined based on the path length d(v') from the initial tree base point to the neighboring modulus point v', and the calculation formula for d(v') is as follows:

[0102] d(v') = min(d(v) + ||v - v'||) (4)

[0103] v' is the nearest neighbor modulo point of v. d(v) and d(v') represent the path lengths from the initial tree base point to the modulo points v and v', and ||v - v'|| represents the Euclidean distance between the modulo points v and v'. In this embodiment, to improve efficiency, only the shortest path between each modulo point and the initial tree base point within a horizontal distance of 10 m is calculated.

[0104] (2) For any modulo point within any of the cuboid voxels, select the shortest path from the path set to obtain the material transportation path of this modulo point, that is, the potential path of this modulo point. The calculation formula for the corresponding initial tree base point is:

[0105] argmin d(s i ) (5)

[0106] where argmin is the index of the initial tree base point corresponding to a modulo point, and d(s i ) is the path length from the modulo point to the initial tree base point s i .

[0107] In another exemplary embodiment of the present application, step 208 specifically includes:

[0108] (1) Determine the initial tree base points whose path quantity exceeds the set threshold as the final tree base points.

[0109] (2) For any of the final tree base points, determine all the material transportation paths passing through this final tree base point as the target path set of this final tree base point.

[0110] (3) For any of the final tree base points, determine the cuboid voxels to which the modulo points corresponding to all the material transportation paths in the target path set belong as the target voxels corresponding to this final tree base point.

[0111] (4) For any of the final tree base points, generate the individual tree segmentation result of this final tree base point according to the point cloud in the target voxel; the individual tree segmentation results of all the final tree base points are used as the individual tree segmentation result of the target forest.

[0112] For the airborne lidar point cloud individual tree segmentation method in the above embodiment, first, the point cloud is organized by cuboid voxels, the near center of gravity point of each voxel is retained as the composition vertex, and the connection edges are generated based on the neighboring voxel vertices, which reduces the number of composition vertices and improves the composition efficiency. Secondly, the constructed connected graph has stronger vertical connectivity, significantly improving the individual tree segmentation accuracy. In addition, the material transportation path theory of the algorithm is less affected by tree species, leaf conditions, understory vegetation, and tree density, and is applicable to a variety of complex forest scenes, ensuring high applicability and accuracy of individual tree segmentation under different conditions. Finally, the cumulative intensity feature of the tree base point is related to the crown volume, which can effectively alleviate the interference of understory vegetation and further improve the segmentation accuracy.

[0113] The following describes an implementation of the method for segmenting individual trees from airborne lidar point clouds and demonstrates the effectiveness of this method.

[0114] The implementation of the method for segmenting individual trees from airborne lidar point clouds includes:

[0115] Step 1: Normalize the height information of the airborne lidar point cloud.

[0116] Step 2: Divide the normalized point cloud into cuboid voxels to form the basic units for extracting model points. The calculation formula for the cuboid voxel is shown in Formula (1) and will not be elaborated here.

[0117] Step 3: In each voxel, extract the laser point closest to the centroid as the model point, providing a basis for constructing the connectivity graph. Among them, the calculation formula for the model point is shown in Formula (2) and will not be elaborated here.

[0118] Step 4: Based on the positions of the model points, connect the model points within their 26-nearest neighbor voxels to generate a connectivity graph to describe the topological relationship between points. Among them, identify the 26-nearest neighbor model points v' of each model point v based on the row, column, and layer numbers of the voxel. The calculation formula is shown in Formula (3) and will not be elaborated here.

[0119] Step 5: Extract the model points of the low-level voxels from the connectivity graph as the initial tree base points to calibrate the potential starting points of the tree trunks.

[0120] Step 6: Using the initial tree base points and the connectivity graph, calculate the shortest paths from the model points to the initial tree base points to form a set of paths for the model points. Among them, the formula for calculating the shortest path from each model point to the initial tree base point using the Dijkstra algorithm is shown in Formula (4) and will not be elaborated here.

[0121] Step 7: Select the shortest paths in the set of paths and define them as the material transportation paths between the model points and the initial tree base points. Among them, after obtaining the potential paths of each model point, retain the shortest path in the set of paths as the material transportation path. The corresponding calculation formula for the initial tree base point is shown in Formula (5) and will not be elaborated here.

[0122] Step 8: According to the material transportation paths, count the number of paths for each initial tree base point and evaluate the cumulative intensity of the initial tree base points.

[0123] Step 9: Screen the initial tree base points and their paths whose number of paths exceeds the set threshold, remove the redundant paths and initial tree base points, and obtain the final tree base points.

[0124] Step 10: Based on the screening results, extract the laser points within the voxels corresponding to the paths to generate the final individual tree segmentation results, and the individual tree segmentation results can be output as a point cloud file (*.las) for more convenient use.

[0125] The above-mentioned airborne lidar point cloud single-tree segmentation method based on the material transportation path realizes the accurate segmentation of single trees by identifying the material transportation path in the point cloud. In the forest point cloud, the starting point and the ending point of the material transportation path correspond to the root point and the crown of the tree respectively. Utilizing this path characteristic, the tree root point and its corresponding crown range are deduced in reverse, thereby completing the single-tree segmentation. Since the material transportation path theory has low dependence on tree species, leaf conditions, understory vegetation, and tree density, this embodiment has good generality and is applicable to various complex forest scenes.

[0126] The effect diagram obtained based on the above steps is as Figure 3 shown. Among them, Figure 3 part (a) shows the lidar point cloud; Figure 3 part (b) shows the construction of the connected graph, where the black points and the gray lines represent the modulus points and edges of the connected graph respectively; Figure 3 part (c) shows the identification of the material transportation path, where the black rectangular frame represents the tree base, and the edges represent the material transportation paths from the modulus points to the tree base, and the edges pointing to the same tree base are coded with the same color; Figure 3 part (d) shows the single-tree segmentation result, where the single-tree point clouds are rendered in different colors, and the gray points represent the understory vegetation.

[0127] Figure 4 This shows an example of the result of the airborne lidar point cloud single-tree segmentation method based on the material transportation path in this embodiment. Compared with existing methods, such as the PCS (Point Cloud Segmentation) algorithm and the AMS3D algorithm, the method in this embodiment shows higher generality in different scenarios and significantly improves the accuracy of single-tree segmentation. The method in this embodiment can achieve high accuracy in forest scenes with different tree species, leaf conditions, understory vegetation, tree density, and point density, and has strong practical value.

[0128] Based on the same inventive concept, the embodiment of the present application also provides an airborne lidar point cloud single-tree segmentation device for implementing the above-mentioned airborne lidar point cloud single-tree segmentation method. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the airborne lidar point cloud single-tree segmentation device provided below can refer to the limitations on the airborne lidar point cloud single-tree segmentation method in the above text, and will not be elaborated here.

[0129] In an exemplary embodiment, as Figure 5 shown, an airborne lidar point cloud single-tree segmentation device is provided, including:

[0130] A point cloud acquisition module 501, configured to acquire a target point cloud; the target point cloud is obtained by scanning a target forest using an airborne lidar.

[0131] A point cloud voxelization module 502, configured to perform point cloud voxelization on the target point cloud to obtain cuboid voxels.

[0132] A mold point extraction module 503, configured to extract mold points within each of the cuboid voxels; the mold point is the laser point closest to the centroid of the cuboid voxel.

[0133] A connected graph generation module 504, configured to generate a connected graph according to the mold points within each of the cuboid voxels and the corresponding neighboring mold point groups; the neighboring mold point groups include: the mold points within the neighboring cuboid voxels of the cuboid voxel where the mold point is located.

[0134] An initial tree base point extraction module 505, configured to extract initial tree base points from the connected graph; the initial tree base points include: the mold points in the low-level voxels; the low-level voxels are the cuboid voxels at a set layer in the connected graph.

[0135] A material transportation path determination module 506, configured to determine the material transportation paths of the mold points within each of the cuboid voxels according to the connected graph and the initial tree base points; the material transportation path is the shortest path in the path set of the mold point; the path set includes: the shortest paths between the mold point and each of the initial tree base points.

[0136] A path number determination module 507, configured to determine the number of material transportation paths passing through any one of the initial tree base points.

[0137] A single tree segmentation result determination module 508, configured to screen the initial tree base points according to the path number to obtain final tree base points, and determine the single tree segmentation result of the target forest according to the final tree base points and the material transportation paths passing through the final tree base points.

[0138] This embodiment is applicable to multi-type forest scenarios with different tree species, leaf states, understory vegetation, tree densities, and point densities, and can significantly improve the accuracy of single-tree scale forest structure characterization, contribute to the in-depth analysis of the complexity and diversity of forest ecosystems, and provide important support for forest management and ecological research.

[0139] In an exemplary embodiment, a computer device is provided. The computer device can be a server or a terminal, and its internal structure diagram can be as Figure 6As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the target point cloud. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a method for single-tree segmentation of airborne lidar point clouds.

[0140] Those skilled in the art can understand that Figure 6 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0141] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0142] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0143] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0144] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memory (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0145] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, etc., without limitation.

[0146] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0147] Specific examples are used in this article to elaborate on the principles and implementation manners of the present application. The descriptions of the above embodiments are only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. An airborne lidar point cloud single tree segmentation method, characterized in that The above-mentioned airborne lidar point cloud single-tree segmentation method includes: Obtaining the target point cloud; the target point cloud is obtained by scanning a target forest using an airborne lidar; Performing point cloud voxelization on the target point cloud to obtain cuboid voxels; Extracting the model points within each of the cuboid voxels; the model points are the laser points closest to the centroid of the cuboid voxel; Generating a connectivity graph based on the model points within each of the cuboid voxels and the corresponding neighboring model point groups; the neighboring model point groups include: the model points within the neighboring cuboid voxels of the cuboid voxel where the model point is located; Extracting the initial tree base points from the connectivity graph; the initial tree base points include: the model points in the low-level voxels; the low-level voxels are the cuboid voxels at a set layer number in the connectivity graph; Determining the material transportation path of the model points within each of the cuboid voxels according to the connectivity graph and the initial tree base points; the material transportation path is the shortest path in the path set of this model point; the path set includes: the shortest paths between this model point and each of the initial tree base points; For any one of the initial tree base points, determining the number of paths of the material transportation path passing through this initial tree base point; Screening the initial tree base points according to the number of paths to obtain the final tree base points, and determining the single-tree segmentation result of the target forest according to the final tree base points and the material transportation paths passing through the final tree base points.

2. The method for segmenting individual trees from airborne lidar point clouds according to claim 1, wherein, Obtaining the target point cloud, specifically including: Collecting the initial lidar point cloud of the target forest using an airborne lidar; Extracting the ground points from the initial lidar point cloud using a cloth simulation filtering algorithm, and constructing a terrain model according to the ground points; Calculating the height difference between the initial lidar point cloud and the terrain model; Taking the height difference as the point cloud elevation, normalizing the height information of the initial lidar point cloud to obtain the height-normalized point cloud, and taking the height-normalized point cloud as the target point cloud.

3. The method for segmenting individual trees from airborne lidar point clouds according to claim 1, wherein, Performing point cloud voxelization on the target point cloud to obtain cuboid voxels, specifically including: Performing point cloud voxelization according to the three-dimensional coordinates of each laser point in the target point cloud to obtain cuboid voxels; the expression of the cuboid voxel is: Among them, (r, c, l) represents the row, column, and layer numbers of the cuboid voxel to which the laser point belongs, where r represents the row number, c represents the column number, and l represents the layer number; int represents the integer function; (x, y, z) represents the three-dimensional coordinates of the laser point, where x represents the x-axis coordinate value, y represents the y-axis coordinate value, and z represents the z-axis coordinate value; x min represents the minimum x-axis coordinate value of the target point cloud, y min represents the minimum y-axis coordinate value of the target point cloud, z min represents the minimum z-axis coordinate value of the target point cloud; v xy represents the horizontal dimension of the cuboid voxel; v z represents the vertical dimension of the cuboid voxel.

4. The method for single-tree segmentation of airborne lidar point clouds according to claim 1, wherein Extracting the model points within each of the cuboid voxels, specifically including: For any one of the cuboid voxels, calculating the Euclidean distance between each laser point within this cuboid voxel and the centroid of this cuboid voxel, and determining the laser point with the minimum Euclidean distance as the model point within this cuboid voxel.

5. The method for segmenting individual trees from airborne lidar point clouds according to claim 1, wherein Generating a connectivity graph based on the model points within each of the cuboid voxels and the corresponding neighboring model point groups, specifically including: For the model points within any one of the cuboid voxels, determining the model points within the 26 neighboring cuboid voxels of the cuboid voxel where the model point is located as the neighboring model point group of this model point; For the model points within any one of the cuboid voxels, connecting this model point and each of the model points in the corresponding neighboring model point group to generate the topological graph of this model point; Determining the connectivity graph according to the topological graphs of the model points within all the cuboid voxels.

6. The method for segmenting individual trees from airborne lidar point clouds according to claim 1, characterized in that, Determining the material transportation path of the model points within each of the cuboid voxels according to the connectivity graph and the initial tree base points, specifically including: For any voxel point in the cuboid voxel, according to the connectivity graph, use the Dijkstra algorithm to calculate the shortest path between the voxel point and each of the initial tree base points, and obtain the path set of the voxel point; For any voxel point in the cuboid voxel, select the shortest path from the path set to obtain the material transportation path of the voxel point.

7. The method for segmenting individual trees from airborne lidar point clouds according to claim 1, wherein Screen the initial tree base points according to the number of paths to obtain the final tree base points, and determine the individual tree segmentation result of the target forest according to the final tree base points and the material transportation paths passing through the final tree base points, specifically including: Determine the initial tree base points with the number of paths exceeding the set threshold as the final tree base points; For any of the final tree base points, determine all the material transportation paths passing through the final tree base point as the target path set of the final tree base point; For any of the final tree base points, determine the cuboid voxel to which the voxel points corresponding to all the material transportation paths in the target path set belong as the target voxel corresponding to the final tree base point; For any of the final tree base points, generate the individual tree segmentation result of the final tree base point according to the point cloud in the target voxel; the individual tree segmentation results of all the final tree base points are used as the individual tree segmentation result of the target forest.

8. An airborne lidar point cloud single-tree segmentation device, characterized in that, The airborne lidar point cloud individual tree segmentation device includes: A point cloud acquisition module for acquiring a target point cloud; the target point cloud is obtained by scanning a target forest using an airborne lidar; A point cloud voxelization module for voxelizing the target point cloud to obtain cuboid voxels; A voxel point extraction module for extracting voxel points in each of the cuboid voxels; the voxel point is the laser point closest to the centroid of the cuboid voxel; A connectivity graph generation module for generating a connectivity graph according to the voxel points in each of the cuboid voxels and the corresponding neighboring voxel point groups; the neighboring voxel point groups include: the voxel points in the neighboring cuboid voxels of the cuboid voxel where the voxel point is located; An initial tree base point extraction module for extracting initial tree base points from the connectivity graph; the initial tree base points include: the voxel points in the low-level voxels; the low-level voxels are the cuboid voxels at a set layer in the connectivity graph; A material transportation path determination module for determining the material transportation paths of the voxel points in each of the cuboid voxels according to the connectivity graph and the initial tree base points; the material transportation path is the shortest path in the path set of the voxel point; the path set includes: the shortest paths between the voxel point and each of the initial tree base points; A path number determination module for determining, for any of the initial tree base points, the number of material transportation paths passing through the initial tree base point; An individual tree segmentation result determination module for screening the initial tree base points according to the number of paths to obtain the final tree base points, and determining the individual tree segmentation result of the target forest according to the final tree base points and the material transportation paths passing through the final tree base points.

9. A computer device, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the airborne lidar point cloud individual tree segmentation method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the airborne lidar point cloud single tree segmentation method described in any one of claims 1-7.

Images