The method of establishing an intelligent tree database through point cloud data

Abstract

This application provides a method for establishing an intelligent tree database using point cloud data, applicable to the field of tree measurement technology. The method includes: acquiring three-dimensional point cloud data; using the three-dimensional point cloud data to represent a forest scene, the point cloud data carrying coordinate information; segmenting the three-dimensional point cloud data into individual tree segments to determine the point cloud model and coordinate information of each tree; measuring each tree based on its point cloud model to obtain its tree structure parameters; and generating a tree database corresponding to the forest scene based on the coordinate information and tree structure parameters of each tree. This application's embodiments can automatically generate a structured tree database including detailed information on individual trees based on three-dimensional point cloud data, shortening tree measurement time, improving the accuracy of tree measurement, enabling large-scale measurement, and thus improving scalability in large-scale forest management or ecological research.

Description

1. Description of a Method for Establishing an Intelligent Tree Database Using Point Cloud Data Technical Field This application relates to the field of tree measurement technology, and more particularly to a method for establishing an intelligent tree database using point cloud data. Background Art Traditional tree measurement involves multiple manual steps, such as geolocation using RTK (Real-time kinematic) measurement under dense canopies. Due to signal obstruction, measuring each tree can take up to approximately 2 minutes. Measurement of tree structure (such as diameter at breast height) requires measuring the trunk circumference with a tape measure and manually converting it to diameter, which often leads to measurement errors. Tree height measurement relies on optical instruments, but accurately locating the treetop under dense canopies is very challenging. Canopy length measurement is more subjective because it relies heavily on human judgment to determine the maximum horizontal distance of the tree, leading to inconsistent results. Furthermore, labeling trees for manual listing adds to the inefficiency. On average, surveying a tree takes about 5 minutes, which greatly limits scalability in large-scale forest management or ecological research. Therefore, traditional tree measurement methods are not only time-consuming but also prone to errors, necessitating a new tree measurement method to provide a scalable and accurate solution for tree surveys and forest management. 20 Summary of the Invention In view of this, embodiments of this application provide a method for establishing an intelligent tree database using point cloud data to solve the problems of time-consuming, error-prone, or unscalable tree measurement methods in the prior art. A first aspect of this application provides a method for establishing an intelligent tree database using point cloud data, comprising: 25 Acquiring three-dimensional point cloud data; the three-dimensional point cloud data is used to represent point cloud data of a forest scene, and the point cloud data carries coordinate information; Segmenting the three-dimensional point cloud data into individual trees to determine the point cloud model and coordinate information of each tree; Measuring each tree based on its point cloud model to obtain the tree structure parameters of each tree; Generating a tree database corresponding to the forest scene based on the coordinate information and tree structure parameters of each tree. In one possible implementation, after segmenting the 3D point cloud data into individual trees and determining the point cloud model and coordinate information of each tree, the process further includes: fitting the canopy of each tree based on the point cloud model and coordinate information of each tree to obtain a graphic identifier corresponding to the location of each tree; numbering each tree to obtain its numbering information; and determining a tree measurement map based on the graphic identifier and numbering information corresponding to the location of each tree.10 In one possible implementation, the 3D point cloud data is segmented into individual trees to determine the point cloud model and coordinate information of each tree, including: extracting seed points based on the 3D point cloud data to determine the seed points of each tree; determining the coordinate information of the seed points of each tree as the coordinate information of each tree; using a preset clustering algorithm to cluster the tree point cloud data to obtain multiple clustered point clouds; 15 Based on the multiple clustered point clouds, determining multiple nodes corresponding to each clustered point cloud; using a preset shortest path algorithm, assigning each node to the corresponding seed point to obtain the point cloud model of each tree. In one possible implementation, the seed point extraction based on the 3D point cloud data to determine the seed points of each tree includes: 20 performing point cloud layering on the 3D point cloud data to determine tree point cloud data and non-tree point cloud data; determining the trunk base point cloud data based on the tree point cloud data and non-tree point cloud data; determining the tree trunk base point cloud data based on the tree trunk base point cloud data. In one possible implementation, the non-tree point cloud data includes ground points and noise points; based on the tree point cloud data and the non-tree point cloud data, the trunk base point cloud data is determined, including: 25. Using a preset filter, ground points are isolated and noise points are removed from the relevant areas of the trunk base in the tree point cloud data to obtain denoised point cloud data; a digital terrain model is generated based on the ground points; in the denoised point cloud data, the digital terrain model is offset, and the point cloud data between the digital terrain model and the offset digital terrain model is segmented and noise points are removed to determine the trunk base point cloud data. 5. In one possible implementation, based on the trunk base point cloud data, the seed point of each tree is determined, including: using a preset clustering algorithm to identify point clusters in the trunk base point cloud data to determine the trunk base point cluster of each tree; the center point of the trunk base point cluster of each tree is used as the seed point of each tree. 10 In one possible implementation, a preset clustering algorithm is used to cluster the tree point cloud data to obtain multiple clustered point clouds, including: cutting the tree point cloud data vertically into multiple first horizontal slices of a first preset thickness; for each first horizontal slice, the preset clustering algorithm is used to cluster the point cloud in the first horizontal slice to obtain multiple clustered point clouds.15 In one possible implementation, after determining multiple nodes corresponding to each of the multiple clustered point clouds, and before assigning each node to a corresponding seed point using a preset shortest path algorithm to obtain the point cloud model of each tree, the method further includes: connecting nodes and seed points with edges to obtain an initial 3D graph; the length of the edges is less than a preset threshold; in the initial 3D graph, removing lines with fewer than a preset number of edges and removing 20 nodes not connected to seed points to obtain a target 3D topology; the lines include a set of nodes and edges; assigning each node to a corresponding seed point using a preset shortest path algorithm to obtain the point cloud model of each tree, including: using a preset shortest path algorithm to determine the distance from each node to all seed points in the target 3D topology; 25 Based on the distance from each node to all seed points, assigning each node to the nearest seed point, HK 30135842 A 4 to obtain the point cloud model of each tree. In one possible implementation, before assigning each node to a corresponding seed point using a preset shortest path algorithm to obtain the point cloud model of each tree, the method further includes: in response to a coordinate input operation, acquiring input coordinate information, and using the point corresponding to the input coordinate information in the 3D point cloud data as a seed point. 5 The second aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method of the first aspect. Compared with the prior art, the embodiments of this application have at least the following technical effects: The method of establishing an intelligent tree database through point cloud data in the first aspect of this application can acquire 3D point cloud data. Since the 3D point cloud data is used to represent the point cloud data of a forest scene, and the point cloud data carries coordinate information, the 3D point cloud data can be segmented into individual trees to determine the point cloud model and coordinate information of each tree; then, based on the point cloud model of each tree, each tree is measured to obtain the tree structure parameters of each tree; furthermore, based on the coordinate information and tree structure parameters of each tree, a tree database corresponding to the forest scene is generated. Therefore, the embodiments of this application can automatically generate a structured tree database including detailed information of individual trees based on 3D point cloud data, avoiding problems such as manual measurement and manual compilation of tree structure parameter lists, shortening the tree measurement time, improving the accuracy of tree measurement, enabling large-scale measurement, and thus improving scalability in large-scale forest management or ecological research. It is understood that the beneficial effects of the second aspect described above can be found in the relevant description in the first aspect above, and will not be repeated here.Brief Description of the Drawings To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 is a flowchart of a method for establishing an intelligent tree database through point cloud data provided in an embodiment of this application; Figure 2 is a flowchart of another method for establishing an intelligent tree database through point cloud data provided in an embodiment of this application; Figure 3 is a flowchart of a method for segmenting three-dimensional point cloud data into single trees and determining the point cloud model and coordinate information of each tree provided in an embodiment of this application; Figure 4 is a schematic diagram of a tree measurement map provided in an embodiment of this application; Figure 5 is a structural schematic diagram of a device for establishing an intelligent tree database through point cloud data provided in an embodiment of this application; Figure 6 is a structural schematic diagram of an electronic device provided in an embodiment of this application. Detailed Description In the following description, specific details such as particular system structures and technologies are set forth for illustration rather than limitation in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application can also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted to avoid unnecessary detail that could obscure the description of this application. It should be understood that, as used in this specification and the appended claims, the term "comprising" indicates the presence of a described feature, integral, step, operation, element, and / or component, but does not exclude the presence or addition of 20 or more other features, integrals, steps, operations, elements, components, and / or a collection thereof. It should also be understood that, as used in this specification and the appended claims, the term "and / or" refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations. In the description of this application, unless otherwise stated, the " / " used in the specification and appended claims 25 indicates that the related objects are in an "or" relationship. For example, A / B can represent A HK 30135842 A 6 or B. The "and / or" in this application merely describes the relationship between the related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Furthermore, in the description of this application, unless otherwise stated, "multiple" means two or more."At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can mean: a, b, c, a and b, a and c, b and c, or a, b, and c. Wherein, a, b, and c can be single or multiple. As used in this specification and the appended claims, the term "if" can be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determination" or "if [the described condition or event] is detected" can be interpreted, depending on the context, as "once determination," "in response to determination," "once [the described condition or event]," or "in response to detection." Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance. References to "one embodiment" or "some embodiments" in this application mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and their variations all mean "including but not limited to," unless otherwise specifically emphasized. The technical solutions of this application and how they solve the aforementioned technical problems are described in detail below with specific embodiments. It should be noted that the following embodiments can be referenced, borrowed from, or combined with each other. Identical terms, similar features, and similar implementation steps in different embodiments will not be repeated. HK 30135842 A 7 Referring to Figure 1, this application embodiment provides a flowchart of a method for establishing an intelligent tree database through point cloud data. As shown in Figure 1, the method for establishing an intelligent tree database through point cloud data in this application embodiment includes steps S101 to S104. S101: Obtain three-dimensional point cloud data; the three-dimensional point cloud data is used to represent point cloud data of a forest scene, and the point cloud data carries coordinate information. 5 Optionally, the three-dimensional point cloud data can be point cloud data obtained by ground-based laser scanning, mobile laser scanning, UAV lidar, airborne laser scanning, etc. For example, staff can carry a mobile laser to walk and scan in a sample plot to obtain the three-dimensional point cloud data of the sample plot, which can be a forest.Optionally, the coordinate information is geographic location information, which can be GPS (global positioning system) coordinates, expressed in latitude and longitude, where N represents North latitude and E represents East longitude. 10 S102. Perform single-tree segmentation on the 3D point cloud data to determine the point cloud model and coordinate information of each tree. In this embodiment, the point cloud model of each tree can be segmented from the 3D point cloud data. Since each point cloud data carries coordinate information, the coordinate information of each tree can be the coordinate information of a specific point cloud data, such as: seed point. 15 S103. Based on the point cloud model of each tree, measure each tree to obtain the tree structure parameters of each tree. Optionally, the tree structure parameters are used to represent the structural information of the tree, and may include: tree height, diameter at breast height, and crown length, etc. S104. Based on the coordinate information and tree structure parameters of each tree, generate a tree database corresponding to the forest scene. Optionally, the tree database includes a list of trees, which can be saved in a tabular format (such as xlsx, csv, etc.). The tree list can list the coordinate information and structural parameters of each tree. The method for establishing an intelligent tree database through point cloud data in this application embodiment can acquire three-dimensional point cloud data. Since the three-dimensional point cloud data is used to represent the point cloud data of the forest scene, the point cloud data carries coordinate information. The three-dimensional point cloud data can be segmented into individual trees to determine the point cloud model and coordinate information of each tree. Then, based on the point cloud model of each tree, each tree is measured to obtain the tree structural parameters of each tree. Furthermore, based on the coordinate information and tree structural parameters of each tree, a tree database corresponding to the forest scene is generated. Therefore, this application embodiment can automatically generate a structured tree database including detailed information of individual trees based on three-dimensional point cloud data, avoiding problems such as manual measurement and manual compilation of tree structural parameter lists, shortening the tree measurement time, improving the accuracy of tree measurement, enabling large-scale measurement, and thus improving the scalability in large-scale forest management or ecological research. In some embodiments, after segmenting the 3D point cloud data into individual trees and determining the point cloud model and coordinate information of each tree, the method further includes: fitting the canopy of each tree based on the point cloud model and coordinate information of each tree to obtain a graphic identifier corresponding to the location of each tree; numbering each tree to obtain its numbering information; and determining a tree measurement map based on the graphic identifier and numbering information corresponding to the location of each tree. This embodiment also solves the problem that traditional methods for establishing intelligent tree databases from point cloud data require manual annotation and organization in the post-measurement processing stage to generate tree plans and tree lists (i.e., tree lists).Research has found that traditional methods of building intelligent tree databases using point cloud data fail to fully utilize the potential of modern technologies, such as 3D point cloud data, which can provide more detailed and accurate information about tree structure. Traditional surveys focus only on basic parameters such as tree height and diameter at breast height (DBH), while point cloud models of individual trees can calculate more advanced metrics, such as tree volume, and further facilitate the estimation of weight²⁰ and biomass, which are crucial for vegetation research and ecological modeling. Against this backdrop, embodiments of this application aim to address these challenges by designing an end-to-end framework to improve the efficiency and accuracy of tree surveys and tree inventory generation. Specifically, embodiments of this application aim to provide: (1) tree measurement maps automatically generated based on 3D point cloud data with accurate geolocation; and (2) a list of trees containing accurate structural parameters of individual trees. However, existing point cloud tree segmentation²⁵ algorithms require extensive parameter tuning and are mostly developed based on temperate forests. Some subtropical regions, such as HK 30135842 A 9, have more complex forest structures, thus requiring the development of an improved tree segmentation algorithm with fewer parameter adjustments and higher scalability for application in complex forest environments. Further research indicates that tree surveys and the generation of tree measurement maps are traditionally tasks requiring manual work by domain experts (such as arborists). However, LiDAR (Light Detection and Ranging) technology can replace these manual processes, automating data collection and analysis with higher accuracy and efficiency. The method for establishing a smart tree database through point cloud data in this application embodiment employs an end-to-end framework as a bridge connecting tree measurement with the GIS environment and large-scale spatial data systems, achieving seamless integration with smart city plans and smart tree management. Referring to Figure 2, this application embodiment provides a flowchart of another method for establishing a smart tree database through point cloud data. As shown in Figure 2, the method for establishing a smart tree database through point cloud data in this application embodiment includes steps S201 to S207. S201. Acquire 3D point cloud data; the 3D point cloud data is used to represent the point cloud data of the forest scene, and the point cloud data carries coordinate information. S202. Perform single-tree segmentation on the 3D point cloud data to determine the point cloud model and coordinate information of each tree. Steps S201 to S202 in this embodiment are consistent with the principles of steps S101 to S102 in this embodiment, and will not be repeated here. S203. Based on the point cloud model and coordinate information of each tree, fit the crown of each tree to obtain a graphic identifier corresponding to the location of each tree. Optionally, the graphic identifier can be a circle or a polygon. S204. Number each tree to obtain the numbering information of each tree. Optionally, the numbering information of each tree can be used in tree measurement maps and tree lists to identify the corresponding tree.S205. Based on the graphic identifier and numbering information corresponding to the location of each tree, determine the tree measurement map. Optionally, the tree measurement map is a tree plan view, which can be represented as a top view of the trees. 25 S206. Based on the point cloud model of each tree, measure each tree to obtain the tree structure parameters of each tree. HK 30135842 A 10 S207. Based on the coordinate information and tree structure parameters of each tree, generate a tree database corresponding to the forest scene. Steps S206 to S207 in this embodiment are consistent with the principles of steps S103 to S104 in this embodiment, and will not be repeated here. 5 Optionally, generating a tree database corresponding to the forest scene based on the coordinate information and tree structure parameters of each tree includes: generating a tree database corresponding to the forest scene based on the coordinate information, tree structure parameters, and numbering information of each tree. In the tree database example of this application embodiment, each tree records its precise HK1980 grid coordinates and corresponding tree structural parameters, including tree height, diameter at breast height (DBH), and crown length. Referring to Figure 3, this application embodiment provides a flowchart for segmenting three-dimensional point cloud data into individual trees to determine the point cloud model and coordinate information of each tree. As shown in Figure 3, segmenting three-dimensional point cloud data into individual trees to determine the point cloud model and coordinate information of each tree includes steps S301 to S305. S301: Extracting seed points based on three-dimensional point cloud data to determine the seed points of each tree. This application embodiment can use a preset seed point extraction algorithm to extract seed points based on three-dimensional point cloud data to determine the seed points of each tree. In some embodiments, seed point extraction based on 3D point cloud data to determine the seed point of each tree includes: performing point cloud layering on the 3D point cloud data to determine tree point cloud data and non-tree point cloud data; 20 determining trunk base point cloud data based on the tree point cloud data and non-tree point cloud data; wherein, the trunk base is the transition area connecting the stem and roots of the tree; and determining the seed point of each tree based on the trunk base point cloud data. Embodiments of this application can layer 3D point cloud data into tree point cloud data and non-tree point cloud data. Non-tree point cloud data includes ground points and noise points, such as grass, shrubs, or human elements.25 In some embodiments, non-tree point cloud data includes ground points and noise points; HK 30135842 A 11 Based on tree point cloud data and non-tree point cloud data, determining the trunk base point cloud data includes: using a preset filter to isolate ground points and remove noise points from the relevant area of ​​the trunk base in the tree point cloud data to obtain denoised point cloud data; generating a digital terrain model based on the ground points; in the denoised point cloud data, offsetting the digital terrain model, and segmenting the point cloud data between the digital terrain model and the offset digital terrain model by 5, and removing noise points to determine the trunk base point cloud data. Optionally, the preset filter can use a Cloth Simulation Filter (CSF), which can isolate ground points and apply a resolution of 0.05. Then, different filters can be used alternately to remove noise points until the relevant area of ​​the trunk base is close to noise-free. 10 As an example, embodiments of this application utilize ground points extracted by CSF to generate a high-resolution digital terrain model (DTM). Then, the DTM is offset upwards by 0.2 meters along the z-axis, and the point cloud between the original DTM and the offset DTM is segmented to obtain the initial trunk base point cloud data. A preset filter is then applied to the initial trunk base point cloud data to remove noise, thus determining the trunk base point cloud data. In the trunk base point cloud data, each point cluster represents a tree. A clustering algorithm can be used to identify each point cluster, and the center point of each point cluster is defined as the seed point of each tree to extract the seed point. In some embodiments, determining the seed point of each tree based on the trunk base point cloud data includes: using a preset clustering algorithm to identify point clusters in the trunk base point cloud data to determine the trunk base point clusters of each tree; and using the center point of each tree's trunk base point cluster as the seed point of each tree. 20 Optionally, the clustering algorithm can be the HDBSCAN (Hierarchical Density-Based Spatial Clustering of Applications with Noise) algorithm, which is a density-based hierarchical clustering algorithm. S302. Determine the coordinate information of the seed point of each tree, as the coordinate information of each tree. Optionally, the center point of the trunk base point group can be a point cloud data of the trunk base point group. Since each point cloud data carries coordinate information, the seed point is determined, and the coordinate information of the seed point can be corresponding to HK 30135842 A 12. S303. Use the preset clustering algorithm to cluster the tree point cloud data to obtain multiple clustered point clouds.In some embodiments, a preset clustering algorithm is used to cluster the tree point cloud data to obtain multiple clustered point clouds, including: 5. Cutting the tree point cloud data vertically into multiple first horizontal slices of a first preset thickness; For each first horizontal slice, the preset clustering algorithm is used to cluster the point cloud in the first horizontal slice to obtain multiple clustered point clouds. Optionally, the first preset thickness can be set according to the actual application. As an example, in this embodiment, the tree point cloud of the sample plot is cut into multiple first horizontal slices along the z-axis, and the thickness of each first horizontal slice is 0.1 meters (i.e., the first preset thickness). Then, the HDBSCAN algorithm is applied to each first horizontal slice to cluster the point cloud data. This can ensure the uniformity of clustering and avoid insufficient clustering in the trunk area and excessive clustering in the crown area. S304. Based on the multiple clustered point clouds, determine multiple nodes corresponding to the multiple clustered point clouds respectively. In some embodiments, based on multiple clustered point clouds, determining up to 15 nodes corresponding to each clustered point cloud includes: for each clustered point cloud, using the center point of the clustered point cloud as a node. Optionally, the center point of the HDBSCAN algorithm cluster is defined as a node, and the generated seed points and nodes convert the unorganized original point cloud into simplified data for easier subsequent processing. Optionally, the center point of each clustered point cloud can be selected from one point cloud in the clustered point cloud. S305, using a preset shortest path algorithm, each node is assigned to the corresponding seed point to obtain a point cloud model of 20 per tree. In some embodiments, after determining multiple nodes corresponding to the multiple clustered point clouds based on multiple clustered point clouds, and before assigning each node to the corresponding seed point using a preset shortest path algorithm to obtain the point cloud model of each tree, the method further includes: connecting nodes and seed points with edges to obtain an initial 3D graph; the length of the edge is less than a preset threshold; in the initial 3D graph, removing lines with fewer than a preset number of edges and removing nodes not connected to seed points to obtain the target 3D topology; the line includes a set of nodes and edges. Optionally, the preset threshold and preset number can be set according to actual application. In this embodiment, nodes and seed points are connected to each other with edges to form an initial 3D graph, i.e., a 3D graph. A pair of points is connected by an edge, and a preset threshold L (meters) is set to limit the maximum length of the edge. The preset threshold aims to avoid over-connection and ensure that the constructed 3D graph can accurately reflect the 3D topology of the sample plot. Optionally, the preset threshold L ranges from 0.4 meters to 0.8 meters. For complex forests such as tropical and mixed-species forests, a lower threshold is recommended; while for planted forests or single-species forests, a higher threshold is suitable. In this application embodiment, a line is defined as a series of edges, containing a set of nodes and edges. Lines with fewer than 10 preset edges (n) are removed.The range of n is from 10 to 30, and the default value of n can be 20 in this embodiment. Removing lines with fewer than a preset number of edges can remove floating point groups that do not belong to the tree structure. In addition, this embodiment also removes nodes that are not connected to the seed point through the graph, so that after removing small sub-components, the final graph is ready for single tree segmentation. 15 Further, each node belongs to a tree component, and the seed point marks the position of each tree. Tree segmentation is achieved by assigning each node to its corresponding seed point. In some embodiments, a preset shortest path algorithm is used to assign each node to the corresponding seed point to obtain the point cloud model of each tree, including: using the preset shortest path algorithm, determining the distance from each node to all 20 seed points in the target three-dimensional topology; based on the distance from each node to all seed points, assigning each node to the nearest seed point to obtain the point cloud model of each tree. Optionally, the shortest path algorithm uses Dijkstra's algorithm, applying the concept of shortest path, and using Dijkstra's algorithm to calculate the shortest path distance from the node to all seed points along the target three-dimensional topology. Based on the 25 Dijkstra distance, nodes are assigned to the nearest seed point. This step is repeated for all nodes, and preliminary tree separation is completed at the node level (HK 30135842 A 14). As an example, this application embodiment demonstrates the logic of a tree segmentation process. Define the target 3D topology as G; where N = set of nodes, = set of edges; 5 The weight of each edge is the Euclidean distance between two points i and j, and its weight can be expressed as w: The shortest path distance from seed point s to node n can be defined as d, which is calculated as: where P = set of connections between s and n; 10 Finally, the node assignment can be defined as: where ⨀ = nearest seed point; = set of seed points; assigned to ⨀; 15 In order to achieve the segmentation of a single tree point cloud, this application embodiment converts the nodes back to HDBSCAN point cloud clustering to reconstruct the real structure of the tree. The overall algorithm only requires input of two main thresholds (L and n), which is more user-friendly and can adapt to various terrain environments in complex forests. In some embodiments, before assigning each node to a corresponding seed point using a preset shortest path algorithm to obtain the point cloud model of each tree, the method further includes: 20. In response to a coordinate input operation, acquiring input coordinate information and using the point corresponding to the input coordinate information in the 3D point cloud data as a seed point. In addition to automated seed point extraction via DTM cutting, this application embodiment also supports manually defined seed points for single-tree segmentation. This method achieves accurate segmentation in extremely complex environments and allows for manual intervention when necessary.This application embodiment considers that when the scanning scene is extremely complex and noisy, the establishment of seed points can incorporate manual intervention. Current single-tree segmentation algorithms are rule-based, relying on predetermined logic to establish seed points. However, when the complexity of the scene exceeds the scope of the applicable logic, manual judgment is required to determine the seed points. This application embodiment can also develop artificial intelligence models to cope with these challenging scenes. Deep learning models can analyze patterns in point clouds, identify complex structures, and autonomously establish accurate seed points, thereby reducing reliance on manual input and improving efficiency in highly complex environments. This application embodiment can avoid over-segmentation or under-segmentation of single trees by establishing seed points during tree segmentation. In addition, when the vegetation is too dense and complex, and there is a lot of noise due to low shrubs and grass, this application embodiment can also input manually determined seed points, demonstrating the flexibility of the algorithm. The single-tree segmentation algorithm of this application embodiment includes seed point extraction, HDBSCAN clustering method, and segmentation method based on Dijkstra's algorithm. The single-tree segmentation algorithm of this application embodiment reduces the need for parameter adjustment and can effectively segment trees in various complex forest types. The main source of error in the single-tree segmentation algorithm is oversegmentation or undersegmentation. Optimized segmentation results can be achieved by adjusting the algorithm parameters. The seed point extraction in this embodiment strictly determines the number of trees to be segmented (i.e., M seed points generate M trees). Furthermore, the seed points in this embodiment are generated from the tree trunk base point cloud data through DTM extraction, making the single-tree segmentation algorithm applicable to all terrains and various environments, including mixed forests, temperate forests, homogeneous plantations, and even wetland forests. Further, this embodiment uses a layer-based clustering method to convert the scattered point cloud into structured nodes. Converting the scattered point cloud into nodes makes the data more ordered. Connecting these nodes with edges transforms the tree point cloud into a "skeleton model" composed of nodes and edges. During the clustering process, the HDBSCAN algorithm is applied to each layer of the point cloud, cutting the entire scene along the z-axis into a first horizontal slice with a thickness of 0.1 meters. This layer-based clustering method ensures detailed node extraction across different vegetation types (from mixed forests to temperate forests), generating nodes that capture fine-grained structural details. By adding edges to these nodes and using Dijkstra's algorithm for tree segmentation, each node is assigned to the nearest seed point, achieving the separation of overlapping canopies and ensuring that overlapping clusters are not misclassified as the same tree. The embodiments of this application mainly include a single-tree segmentation algorithm and a single-tree measurement algorithm. After obtaining the point cloud model of a single tree from the sample plot, the parameters of each tree can be measured, including tree height, diameter at breast height (DBH), and canopy length.5. In some embodiments, tree structural parameters include at least one of tree height, diameter at breast height (DBH), and crown length; Based on the point cloud model of each tree, measurements are performed on each tree to obtain the tree structural parameters of each tree, including: For the point cloud model of each tree, performing at least one of the following measurement operations: Determining the maximum and minimum values ​​of all coordinate information corresponding to the point cloud model in the vertical direction, and determining the tree height based on the difference between the maximum and minimum values; Fitting a DBH circle at a predetermined height in the point cloud model, and using the diameter of the DBH circle as the DBH; Cutting the point cloud model vertically into multiple second horizontal slices of a second predetermined thickness, and using the distance between the two farthest points in the second horizontal slice with the largest area as the crown length. 15. Optionally, the predetermined height and the second predetermined thickness can be set according to actual application. This application embodiment uses a single-tree measurement algorithm to perform measurement operations on each tree to obtain the tree structural parameters of each tree as an example. The tree height is measured by calculating the difference between the maximum and minimum values ​​of the z-axis of the tree point cloud. The diameter at breast height (DBH) can be defined as the diameter of a 20-DBH circle fitted to the trunk at a point 1.3 to 1.32 meters above the ground. As for the crown length, it can be defined as the distance between the two farthest points in the second horizontal slice, where the point cloud of a single tree is cut into second horizontal slices with a thickness of 0.1 meters. Referring to Figure 4, this embodiment provides a schematic diagram of a tree survey map. As shown in Figure 4, the tree survey map is a top-down plan view of the trees, with each tree corresponding to a circle, and each circle corresponding to a number 25. Figure 4 shows 14 trees, corresponding to numbers 1 to 14. The tree plan view can be associated with the tree database (HK 30135842 A 17) through the number information of each tree. In practical applications, this embodiment can also fit white circles onto the crown of each tree based on the point cloud model of each tree. Each white circle represents the precise geographical location of the tree on the tree survey plan. The digital terrain model (DTM) of the forest scene is set to a white-gray background, and the DTM is generated from ground points in the point cloud. The single-tree measurement algorithm of this embodiment achieves a consistent and accurate survey of tree structure parameters, avoiding errors from manual measurement. The geographic location and canopy boundary of each segmented tree are visualized in a two-dimensional tree measurement map, which is generated from the segmented tree point cloud data and provided as vector data (Shapefile format). Each polygon in the tree measurement map represents a tree canopy. In summary, this embodiment proposes an automated end-to-end framework for generating a detailed tree database from three-dimensional point cloud data. The input of this embodiment is raw point cloud data (e.g., point cloud data of a forest scene), and the output is a structured database containing detailed information about individual trees.Furthermore, this application embodiment achieves automatic drawing of tree measurement maps based on point cloud applications, successfully transforming this previously human-dominated activity into an algorithm-based framework. By automating and standardizing the tree measurement map drawing process, accuracy is improved. Further, this application embodiment can employ a single-tree segmentation algorithm to segment 3D point cloud data into individual trees. This algorithm can separate the point cloud model of each tree from the input 3D point cloud data. This segmentation method can accurately identify each tree in complex environments with overlapping trees. Subsequently, this application embodiment employs a single-tree measurement algorithm to extract key parameters, including tree height, diameter at breast height (DBH), and crown length. In addition, the geographical location of each tree can be measured with high precision, thereby achieving spatial integration with a Geographic Information System (GIS). This application embodiment can be written in Python. This end-to-end framework supports inputting ground or moving point cloud data and outputting a list of trees (in tabular format, such as xlsx, csv, etc.) and tree measurement maps (in Shapefile format, each graphic representing a single tree crown). This application automates the traditional tree survey process, which relies on measuring tapes, on-site tree surveying, and manual drawing of tree plans. The entire process simplifies the traditionally labor-intensive and error-prone task of generating tree lists, providing a scalable and accurate solution for tree surveys and forest management. This application can generate output formats that meet industry or client-specific needs, such as CAD (Calculator-Aided Design). This ensures seamless integration with existing workflows and improves compatibility with industry-standard tools, thereby promoting efficient communication and collaboration between projects. The method of building a smart tree database using point cloud data in this application can also be applied to the following aspects: Tree assessment in construction projects: In construction, tree assessment is typically used to evaluate trees on planned development land. This application can replace the traditional manual tree assessment process by automatically generating tree plans from LiDAR data. Carbon sequestration research: This application can be used to study carbon sequestration in forests by analyzing tree structure and biomass. By acquiring point cloud models of individual trees, the volume of individual trees can be measured, and their carbon storage capacity estimated. Furthermore, the embodiments of this application are applicable to various forest types, from temperate to tropical forests. Tree management: The embodiments of this application can be used to create a detailed tree database and map tree species, sizes, and distribution. This data is crucial for sustainable forestry, biodiversity management, and long-term monitoring of tree health.It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. Referring to Figure 5, this application embodiment provides a schematic diagram of the structure of a device 50 for establishing an intelligent tree database 20 through point cloud data. As shown in Figure 5, the device 50 for establishing an intelligent tree database through point cloud data includes: an acquisition module 501, a first determination module 502, a measurement module 503, and a first generation module 504. The acquisition module 501 is used to acquire three-dimensional point cloud data; the three-dimensional point cloud data is used to represent the point cloud data of the forest scene, and the point cloud data carries coordinate information; the first determination module 502 is used to segment the three-dimensional point cloud data into individual trees, and determine the point cloud model and coordinate information of each tree; the measurement module 503 is used to measure each tree based on the point cloud model of each tree, and obtain the tree structure parameters of each tree; the first generation module 504 is used to generate a tree database corresponding to the forest scene based on the coordinate information and tree structure parameters of each tree. Optionally, the device 50 for establishing an intelligent tree database through point cloud data further includes: a second determination module, a numbering module, and a second generation module. The second determining module is used to fit the canopy of each tree based on the point cloud model and coordinate information of each tree, and obtain the graphic identifier corresponding to the location of each tree; the numbering module is used to number each tree, and obtain the numbering information of each tree; the second generating module is used to determine the tree measurement map based on the graphic identifier and numbering information corresponding to the location of each tree. Optionally, the first determining module is used to extract seed points based on the three-dimensional point cloud data, and determine the seed points of each tree; determine the coordinate information of the seed points of each tree, and use them as the coordinate information of each tree; use a preset clustering algorithm to cluster the tree point cloud data, and obtain multiple clustered point clouds; based on the multiple clustered point clouds, determine multiple nodes corresponding to the multiple clustered point clouds respectively; use a preset shortest path algorithm to assign each node to the corresponding seed point, and obtain the point cloud model of each tree. Optionally, the first determining module is used to perform point cloud layering on the 3D point cloud data to determine tree point cloud data and non-tree point cloud data; based on the tree point cloud data and non-tree point cloud data, to determine the trunk base point cloud data; and based on the trunk base point cloud data, to determine the seed point for each tree. Optionally, the first determining module is used to use a preset filter to isolate ground points and remove noise points from the relevant areas of the trunk base in the tree point cloud data to obtain denoised point cloud data; based on the ground points, to generate a digital terrain model; in the denoised point cloud data, to offset the digital terrain model, and to segment the point cloud data between the digital terrain model and the offset digital terrain model to remove noise points, thereby determining the trunk base point cloud data.25 Optionally, the first determining module is used to perform point group identification on the tree trunk base point cloud data using a preset clustering algorithm, and determine the tree trunk base point group for each tree; the center point of the tree trunk base point group for each tree is used as the seed point for each tree. Optionally, the first determining module is used to cut the tree point cloud data vertically into multiple first horizontal slices of a first preset thickness; for each first horizontal slice, the preset clustering algorithm is used to group the point cloud in the first horizontal slice to obtain multiple clustered point clouds. 5 Optionally, the first determining module is used to use the center point of each clustered point cloud as a node. Optionally, the first determining module is used to connect nodes and seed points through edges to obtain an initial 3D graph; the length of the edge is less than a preset threshold; in the initial 3D graph, lines with fewer than a preset number of edges are removed, and nodes not connected to seed points are removed to obtain a target 3D topology; the line includes a set of nodes and edges; using a preset shortest path algorithm, the distance from each node to all seed points in the target 3D topology is determined; based on the distance from each node to all seed points, each node is assigned to the nearest seed point to obtain a point cloud model of each tree. Optionally, the device 50 for establishing an intelligent tree database through point cloud data further includes: an input module, which is used to respond to coordinate input operations, acquire input coordinate information, and use the points corresponding to the input coordinate information in the 3D point cloud data as seed points. Optionally, the tree structural parameters include at least one of tree height, diameter at breast height (DBH), and crown length. Measurement module 503 performs at least one of the following measurement operations on the point cloud model of each tree: determining the maximum and minimum values ​​of all coordinate information corresponding to the point cloud model in the vertical direction; determining the tree height based on the difference between the maximum and minimum values; fitting a DBH circle at a predetermined height in the point cloud model, using the diameter of the DBH circle as 20 DBH; cutting the point cloud model vertically into multiple second horizontal slices of a second predetermined thickness, using the distance between the two farthest points in the second horizontal slice with the largest area as the crown length. In application, the modules in the device 50 for establishing an intelligent tree database through point cloud data can be software program modules, or they can be implemented through different logic circuits integrated in a processor, or they can be implemented through up to 25 distributed processors.HK 30135842 A 21 The apparatus 50 for establishing a smart tree database through point cloud data in the embodiments of this application can execute the method provided in the embodiments of this application. Their implementation principles are similar. The actions performed by each module in the apparatus 50 for establishing a smart tree database through point cloud data in each embodiment of this application correspond to the steps in the methods of each embodiment of this application. For detailed functional descriptions of each module of the apparatus 50 for establishing a smart tree database through point cloud data, please refer to the descriptions in the corresponding methods shown above; these will not be repeated here. This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the methods in the embodiments of this application. Referring to Figure 6, this application provides a schematic diagram of the structure of an electronic device 60. As shown in Figure 610, the electronic device 60 in this application includes: a memory 62, a processor 61, and a computer program 63 stored in the memory 62 and executable on the processor 61. When the processor 61 executes the computer program, it implements the steps of the methods in the embodiments of this application. The electronic device 60 can be a desktop computer, a laptop, a handheld computer, or a cloud server, etc. Electronic device 60 may include, but is not limited to, processor 61 and memory 62. Those skilled in the art will understand that FIG6 is merely an example of electronic device 60 and does not constitute a limitation on electronic device 60. It may include more or fewer components than illustrated, or combine certain components, or different components, such as input / output devices, network access devices, etc. Processor 61 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor. Memory 62 may be an internal storage unit in some embodiments, such as a hard disk or RAM. The memory 25 62 can be a removable / non-removable, volatile / non-volatile computer system storage medium, for example: memory 62 is a non-volatile memory used for reading and writing non-volatile magnetic media.In other embodiments, memory 62 can also be an external storage device, such as a plug-in hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on electronic device 60. Memory 62 is used to store operating system, application programs, bootloader, data, and other programs, such as program code of computer programs. Memory 62 can also be used to temporarily store data that has been output or will be output. It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application, and their specific functions and technical effects can be found in the method embodiments section, and will not be repeated here. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, only the division of the above 10 functional units and modules is illustrated by example. In actual applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. This application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps in the above method embodiments. If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the implementation of all or part of the processes in the methods of the above embodiments can be accomplished by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, it can implement the steps of the above method embodiments. The aforementioned computer program includes computer program code, which may be in the form of source code, object code, executable file, or some intermediate form.The aforementioned computer-readable medium may include at least: any entity or device capable of carrying computer program code to a device / terminal equipment, recording media, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. Those skilled in the art will understand that implementing all or part of the processes in the above embodiments can be accomplished by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium. When executed, the program can include the processes of the embodiments of the above methods. In this case, the storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc. The storage medium may also include combinations of the above types of memory. This application provides a computer program product that, when run on a processor 15, enables the processor to implement the steps described in the various method embodiments above. In the above embodiments, the descriptions of each embodiment have different focuses; parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware 20. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application. In the embodiments provided in this application, it should be understood that the disclosed apparatus / network devices and methods can be implemented in other ways. For example, the device / network equipment embodiments described above are merely illustrative. For instance, the division of modules or units described above is only a logical functional division; in actual implementation (HK 30135842 A 24), there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings or direct couplings or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.The units described above as separate components may or may not be physically separate. Similarly, the components shown as units 5 may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs. The above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application. HK 30135842 A 1 Claim 1. A method for establishing an intelligent tree database through point cloud data, characterized in that it includes: acquiring three-dimensional point cloud data; the three-dimensional point cloud data is used to represent point cloud data of a forest scene, and the point cloud data carries coordinate information; segmenting the three-dimensional point cloud data into individual trees to determine the point cloud model and coordinate information of each tree; measuring each tree based on the point cloud model of each tree to obtain the tree structure parameters of each tree; and generating a tree database corresponding to the forest scene based on the coordinate information and the tree structure parameters of each tree. 10 2. The method for establishing an intelligent tree database through point cloud data according to claim 1, characterized in that, after segmenting the three-dimensional point cloud data into individual trees and determining the point cloud model and coordinate information of each tree, it further includes: fitting the crown of each tree based on the point cloud model and coordinate information of each tree, 15 to obtain the graphic identifier corresponding to the location of each tree; numbering each tree to obtain the numbering information of each tree; and determining the tree measurement map based on the graphic identifier corresponding to the location of each tree and the numbering information. 3. The method for establishing an intelligent tree database through point cloud data according to claim 1, characterized in that, the step of segmenting the three-dimensional point cloud data into individual trees and determining the point cloud model and coordinate information of each tree includes: extracting seed points based on the three-dimensional point cloud data to determine the seed points of each tree; determining the coordinate information of the seed points of each tree as the coordinate information of each tree; using a preset clustering algorithm to cluster the tree point cloud data to obtain multiple clustered point clouds; HK 30135842 A 2 Based on the multiple clustered point clouds, determining multiple nodes corresponding to the multiple clustered point clouds respectively; using a preset shortest path algorithm to assign each node to the corresponding seed point to obtain the point cloud model of each tree.4. The method for establishing an intelligent tree database through point cloud data according to claim 3, characterized in that, the step of extracting seed points based on the three-dimensional point cloud data to determine the seed points of each tree includes: performing point cloud layering on the three-dimensional point cloud data to determine tree point cloud data and non-tree point cloud data; determining trunk base point cloud data based on the tree point cloud data and the non-tree point cloud data; and determining the seed points of each tree based on the trunk base point cloud data. 10 5. The method for establishing an intelligent tree database through point cloud data according to claim 4, characterized in that the non-tree point cloud data includes ground points and noise points; the step of determining the trunk base point cloud data based on the tree point cloud data and the non-tree point cloud data includes: 15 using a preset filter to isolate the ground points and remove the noise points from the relevant area of ​​the trunk base in the tree point cloud data to obtain denoised point cloud data; generating a digital terrain model based on the ground points; in the denoised point cloud data, offsetting the digital terrain model, and segmenting the point cloud data between the digital terrain model and the offset digital terrain model to remove noise points, 20 determining the trunk base point cloud data. 6. The method for establishing an intelligent tree database through point cloud data according to claim 4, characterized in that, determining the seed point of each tree based on the trunk base point cloud data includes: using a preset clustering algorithm to identify point clusters in the trunk base point cloud data, and determining the trunk base point clusters of each tree; taking the center point of the trunk base point cluster of each tree as the seed point of each tree. HK 30135842 A 3 7. The method for establishing an intelligent tree database through point cloud data according to claim 3, characterized in that, using a preset clustering algorithm to cluster the tree point cloud data to obtain multiple clustered point clouds includes: cutting the tree point cloud data vertically into multiple first horizontal slices of a first preset thickness; for each first horizontal slice, using a preset clustering algorithm to cluster the point cloud in the first horizontal slice to obtain multiple clustered point clouds.8. The method for establishing an intelligent tree database through point cloud data according to claim 3, characterized in that, after determining multiple nodes corresponding to the multiple clustered point clouds based on the multiple clustered point clouds, and before assigning each node to a corresponding seed point using a preset shortest path algorithm to obtain a point cloud model of each tree, the method further includes: connecting the nodes and the seed points with edges to obtain an initial three-dimensional graph; the length of the edge is less than a preset threshold; removing lines with fewer than a preset number of edges in the initial three-dimensional graph, and removing nodes not connected to the seed points to obtain a target three-dimensional topology; the lines include a set of nodes and edges; assigning each node to a corresponding seed point using a preset shortest path algorithm to obtain a point cloud model of each tree includes: using a preset shortest path algorithm to determine the distance from each node to all seed points in the target three-dimensional topology; assigning each node to the nearest seed point based on the distance from each node to all seed points to obtain a point cloud model of each tree. 9. The method for establishing an intelligent tree database through point cloud data according to claim 3, characterized in that, before assigning each node to a corresponding seed point using a preset shortest path algorithm to obtain the point cloud model of each tree, it further includes: HK 30135842 A 4 responding to a coordinate input operation, acquiring input coordinate information, and using the point corresponding to the input coordinate information in the three-dimensional point cloud data as a seed point. 10. An electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 9. HK 30135842 A 1 Specification Drawings Figure 1 HK 30135842 A 2 Figure 2 HK 30135842 A 3 Figure 3 HK 30135842 A 4 Figure 4 HK 30135842 A 5 Figure 5 Figure 6 HK 30135842 A.