Tree surveying method and device based on lidar point cloud data, storage medium and electronic equipment
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GUIZHOU ZHILIN ZHONGCHENG ENGINEERING CONSULTING CO LTD
- Filing Date
- 2024-05-08
- Publication Date
- 2026-06-30
AI Technical Summary
Existing technologies using lidar data collected by drones have low accuracy in identifying trees.
By acquiring radar point cloud data of trees within the target area, target parameters such as tree height, canopy closure, and age group are determined and input into the target model. Using multiple tree species corresponding to diameter at breast height (DBH) identification sub-models and DBH normal distribution functions, the tree species are identified and their rationality is verified.
This greatly improves the accuracy of forest species identification based on lidar data collected by drones.
Smart Images
Figure CN118537388B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of forest tree identification technology, and in particular to a method, apparatus, storage medium and electronic equipment for individual tree survey based on lidar point cloud data. Background Technology
[0002] With the advancement of forestry informatization and intelligentization, drones are playing an increasingly important role in forestry survey data collection, transmission, and analysis. Drones, equipped with remote sensing and aerial photography technologies, are contributing to forestry development, enabling them to "go where humans cannot go, see where humans cannot see, and do what humans cannot do" in forestry surveys. This significantly improves work efficiency and data accuracy while substantially reducing the risks of fieldwork.
[0003] Among related technologies, precise GPS positioning is achieved through data collection using lidar from unmanned aerial vehicles (UAVs). High-definition cameras equipped on the UAVs are then used to input the collected information into a forest resource survey platform, enabling real-time monitoring of the actual situation in forest areas, natural disaster occurrences, and timber harvesting, thereby improving the overall capacity for forestry management.
[0004] However, the identification of trees based on lidar data collected by drones has a technical problem of low accuracy. Summary of the Invention
[0005] The purpose of this application is to provide a method, apparatus, storage medium, and electronic device for single-tree survey based on lidar point cloud data, which can improve the accuracy of tree species identification based on lidar data collected by UAVs.
[0006] This application provides a single-tree survey method based on lidar point cloud data, including:
[0007] The system acquires radar point cloud data of trees within a target area and determines target parameters for each tree within the target area based on the radar point cloud data. The target parameters include tree height, canopy closure, and age group. The target parameters for each tree are input into a target model to determine the tree species to which each tree belongs. The target model includes multiple diameter-at-breast-width (DBH) identification sub-models corresponding to multiple tree species, one DBH sub-model per tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model. The DBH identification sub-models are used to identify the DBH of each tree. The DBH normal distribution function is used to verify the reasonableness of the DBH identified by the DBH identification sub-models.
[0008] Optionally, determining the tree height and canopy closure of each tree within the target area based on the radar point cloud data includes: analyzing the radar point cloud data, analyzing the elevation value of each point and its distance from other points using a point cloud segmentation algorithm, and determining multiple sub-point clouds based on the elevation value of each point and its distance from other points; each sub-point cloud corresponds to one tree; analyzing each of the multiple sub-point clouds to determine the tree height and canopy closure of the tree corresponding to each sub-point cloud, and determining the age group of each tree within the target area by querying afforestation data.
[0009] Optionally, before inputting the target parameters of each tree into the target model and determining the tree species to which each tree belongs, the method further includes: obtaining the site conditions of the target area and determining the normal distribution constraints corresponding to each diameter at breast height (DBH) identification sub-model based on the site conditions; randomly generating the expected value and variance of each DBH identification sub-model based on the normal distribution constraints; and constructing the normal distribution function of DBH corresponding to each DBH identification sub-model based on the expected value and variance of each DBH identification sub-model.
[0010] Optionally, the step of inputting the target parameters of each tree into the target model to determine the tree species to which each tree belongs includes: inputting the target parameters corresponding to the target tree into the diameter at breast height (DBH) identification sub-model corresponding to each tree species to obtain multiple candidate DBH values; one DBH identification sub-model outputs one candidate DBH value; based on the DBH normal distribution function corresponding to each DBH identification sub-model, calculating the target difference between the candidate DBH value corresponding to each DBH identification sub-model and the median of the normal distribution function corresponding to each DBH identification sub-model; selecting the DBH identification sub-model corresponding to the candidate DBH value with the smallest target difference from the multiple candidate DBH values and determining it as the target DBH identification sub-model, and determining the tree species corresponding to the target DBH identification sub-model as the tree species to which the target tree belongs.
[0011] Optionally, the step of inputting the target parameters corresponding to the target trees into the diameter at breast height (DBH) identification sub-model corresponding to each tree species to obtain multiple candidate DBH values includes:
[0012] The diameter at breast height (DBH) identification sub-model for each tree species is constructed based on the following formula:
[0013] DBH = aH^a1[cc]^a2 S^a3 N^a4 (Formula 1)
[0014] Where DBH is diameter at breast height (DBH), H is tree height, cc is canopy closure, S is site conditions, N is tree age obtained based on age group, and a, a1, a2, a3, and a4 are coefficients in the coefficient group, with different coefficient groups corresponding to different tree species.
[0015] Optionally, after inputting the target parameters of each tree into the target model and determining the tree species to which each tree belongs, the method further includes: calculating the number of trees corresponding to each tree species in the target area based on the tree species to which each tree belongs, and calculating the timber yield of each tree based on the tree height and diameter at breast height of each tree.
[0016] This application also provides a single-tree survey device based on lidar point cloud data, comprising:
[0017] The acquisition module is used to acquire radar point cloud data of trees within a target area and determine the target parameters of each tree within the target area based on the radar point cloud data. The target parameters include tree height, canopy closure, and age group. The calculation module is used to input the target parameters of each tree into the target model to determine the tree species to which each tree belongs. The target model includes: multiple diameter at breast height (DBH) identification sub-models corresponding to multiple tree species, one DBH sub-model corresponding to one tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model. The DBH identification sub-model is used to identify the DBH of each tree. The DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-model.
[0018] Optionally, the acquisition module is specifically used to analyze the radar point cloud data, analyze the elevation value of each point and the distance between each point and other points through a point cloud segmentation algorithm, and determine multiple sub-point clouds based on the elevation value of each point and the distance between each point and other points; each sub-point cloud corresponds to a tree; the acquisition module is further used to analyze each of the multiple sub-point clouds, determine the tree height and canopy closure of the tree corresponding to each sub-point cloud, and determine the age group of each tree in the target area by querying afforestation data.
[0019] Optionally, the device further includes: a construction module; the acquisition module is further configured to acquire the site conditions of the target area and determine the normal distribution constraints corresponding to each diameter at breast height (DBH) identification sub-model based on the site conditions; the construction module is configured to randomly generate the expected value and variance of each DBH identification sub-model based on the normal distribution constraints corresponding to each DBH identification sub-model; the construction module is further configured to construct the normal distribution function of DBH corresponding to each DBH identification sub-model based on the expected value and variance of each DBH identification sub-model.
[0020] Optionally, the calculation module is specifically used to input the target parameters corresponding to the target forest tree into the diameter at breast height (DBH) identification sub-model corresponding to each tree species to obtain multiple candidate DBH values; each DBH identification sub-model outputs a candidate DBH value; the calculation module is further used to calculate the target difference between the candidate DBH value corresponding to each DBH identification sub-model and the median of the normal distribution function corresponding to each DBH identification sub-model based on the DBH normal distribution function corresponding to each DBH identification sub-model; the calculation module is further used to select the DBH identification sub-model corresponding to the candidate DBH value with the smallest target difference from the multiple candidate DBH values and determine it as the target DBH identification sub-model, and determine the tree species corresponding to the target DBH identification sub-model as the tree species to which the target forest tree belongs.
[0021] Optionally, the diameter at breast height (DBH) identification sub-model for each tree species is constructed based on the following formula:
[0022] DBH = aH a1 cc a2 S a3 N a4 (Formula 1)
[0023] Where DBH is diameter at breast height (DBH), H is tree height, cc is canopy closure, S is site conditions, N is tree age obtained based on age group, and a, a1, a2, a3, and a4 are coefficients in the coefficient group, with different coefficient groups corresponding to different tree species.
[0024] Optionally, the calculation module is also used to calculate the number of trees of each tree species in the target area based on the tree species to which each tree belongs, and to calculate the timber yield of each tree based on the tree height and diameter at breast height of each tree.
[0025] This application also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the single-tree survey method based on lidar point cloud data as described above.
[0026] 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, wherein the processor executes the program to implement the steps of any of the above-described single-tree survey methods based on lidar point cloud data.
[0027] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the single-tree survey method based on lidar point cloud data as described above.
[0028] The method, apparatus, storage medium, and electronic equipment for single-tree survey based on lidar point cloud data provided in this application first acquire lidar point cloud data of trees within a target area, and then determine the target parameters for each tree within the target area based on the lidar point cloud data. The target parameters include tree height, canopy closure, and age group. Next, the target parameters for each tree are input into a target model to determine the tree species to which each tree belongs. The target model includes multiple diameter-at-breast-width (DBH) identification sub-models corresponding to multiple tree species, with one DBH sub-model per tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model. The DBH identification sub-models are used to identify the DBH of each tree; the DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-models. In this way, the tree species can be accurately identified through the target model, greatly improving the accuracy of tree species identification based on lidar data collected by UAVs. Attached Figure Description
[0029] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0030] Figure 1 This is a flowchart illustrating the single-tree survey method based on lidar point cloud data provided in this application;
[0031] Figure 2 This is a yield comparison table provided in this application;
[0032] Figure 3 This is a schematic diagram of the single-tree survey device based on lidar point cloud data provided in this application;
[0033] Figure 4 This is a schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation
[0034] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0035] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.
[0036] The following describes the technical terms used in the embodiments of this application:
[0037] Canopy closure: Canopy closure refers to the degree to which the canopies of trees in a forest cover the ground, reflecting the density of the forest stand and the degree of light energy utilization. Canopy closure is expressed as the ratio of the vertical projection area of the tree canopies to the forest floor area, usually in decimals or percentages. A canopy closure of 1 or 100% represents complete ground coverage, decreasing sequentially below that. Canopy closure is one of the important indicators of forest structure and a key indicator for controlling logging intensity and classifying forest land types.
[0038] Age groups: Age groups refer to the classification of forest stands according to the growth and development stage of the trees. They are generally divided into five age groups: young forests, middle-aged forests, near-mature forests, mature forests, and over-mature forests. Different age groups require different management practices and utilization methods.
[0039] Elevation value: refers to the height of a point relative to a reference surface, usually expressed in meters. Different reference surfaces have different elevation systems, such as orthographic height, normal height, force height, and geodetic height. Methods for determining elevation values include leveling, trigonometric leveling, and physical leveling.
[0040] Site conditions refer to the comprehensive natural environmental factors related to forest growth and development on afforestation land. This manifests as differences in microclimate, soil, hydrology, vegetation, and other environmental factors across different afforestation plots due to their varying topographical locations. Site conditions are a crucial factor influencing forest classification, management, and afforestation.
[0041] Site conditions mainly include the following five environmental factors:
[0042] Topography: This includes altitude, aspect, shape, gradient, and micro-topography, which affect forest conditions such as sunlight, temperature, moisture, and wind. Soil: This includes soil type, soil layer thickness, humus layer thickness and content, soil erosion, texture, structure, compaction, pH value, gravel content, parent material type, and weathering degree, which affect forest conditions such as nutrients, moisture, and air. Hydrology: This includes groundwater depth and seasonal variations, groundwater mineralization and salinity, the presence and duration of seasonal flooding, and the likelihood of flooding, which affect forest conditions such as moisture and salinity. Biology: This includes the distribution of plant species, species cover, abundance and dominant species, community type, and pest and disease status, which affect forest conditions such as competition, symbiosis, and resilience. Human activities: This includes the historical evolution and current status of land use, and the effects of various human activities on the above environmental factors, which affect forest disturbance, damage, and protection.
[0043] Drone-based LiDAR data acquisition constructs orthophotos and 3D images. Drone imagery and LiDAR enable more precise data collection. The integration of drones with remote sensing technology achieves accurate GPS positioning. High-definition cameras on drones input the collected information into a forest resource survey platform, allowing for real-time monitoring of forest conditions, natural disasters, and logging activity, thus improving comprehensive forestry management capabilities. 3D imagery enables separate office areas, eliminating the need for staff to traverse mountain boundaries, and is suitable for comprehensive perception of complex scenes with large-scale, high-precision, and high-definition capabilities. Combined with previous forest reform data, boundary demarcation is completed. This efficient data acquisition equipment and professional data processing workflow generate data results that intuitively reflect the appearance, location, height, and other attributes of terrain features, ensuring realistic results and surveying-grade accuracy.
[0044] The drone's LiDAR provides 3D information of the entire forest. Based on this 3D information, the following steps are taken: Point cloud elevation normalization is performed on the classified point cloud data to avoid the influence of terrain undulations on tree elevation and better represent the actual height of the trees. A distance-based clustering algorithm is then used to segment individual trees from the elevation-normalized point cloud. This algorithm uses the elevation of the analyzed points and the distance between points as the criteria, allocating points from top to bottom based on a spacing threshold and a minimum spacing. The crown vertex of each tree is treated as a separate cluster, and the tree height is extracted by progressively classifying points by comparing the distances from other points below the vertex to the vertex.
[0045] To address the technical problem of low accuracy in individual tree survey methods in related technologies, this application provides an individual tree survey method based on lidar point cloud data. This method outputs multiple diameters at breast height (DBH) through multiple DBH sub-models in the target model. Then, the reasonableness of the DBH output by the DBH sub-model is verified by the DBH normal distribution function corresponding to each DBH sub-model, thereby accurately identifying the tree species. This greatly improves the accuracy of tree species identification based on lidar data collected by UAVs.
[0046] The single-tree survey method based on lidar point cloud data provided in this application will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.
[0047] like Figure 1 As shown in the embodiment of this application, a single-tree survey method based on lidar point cloud data is provided. The method may include the following steps 101 and 102:
[0048] Step 101: Obtain radar point cloud data of trees within the target area, and determine the target parameters of each tree within the target area based on the radar point cloud data.
[0049] The target parameters include: tree height, canopy closure, and age group.
[0050] For example, the aforementioned radar point cloud data can be obtained by scanning the target area using a lidar mounted on a drone. The aforementioned canopy closure can be the canopy closure of the statistical area to which the trees within the target area belong; that is, within the statistical area, the canopy closure in the target parameters of each tree can be the same. Depending on the size of the statistical area and the size of the target area, the target area can include one or more statistical areas.
[0051] Specifically, step 101 above may further include the following steps 101a1 and 101a2:
[0052] Step 101a1: Analyze the radar point cloud data, analyze the elevation value of each point and the distance between each point and other points using a point cloud segmentation algorithm, and determine multiple sub-point clouds based on the elevation value of each point and the distance between each point and other points.
[0053] Each sub-point cloud corresponds to a tree.
[0054] Step 101a2: Analyze each of the multiple sub-point clouds to determine the tree height and canopy closure of the corresponding trees in each sub-point cloud, and determine the age group of each tree in the target area by querying afforestation data.
[0055] For example, the lidar on the drone is used to measure the trees in the target area multiple times to obtain a large amount of point cloud data; the point cloud segmentation algorithm is used to segment the trees in the point cloud according to the elevation value and distance of the points; the attribute extraction algorithm is used to calculate the location, tree height, canopy closure and other information of each tree according to the shape and size of the trees in the point cloud.
[0056] Step 102: Input the target parameters of each tree into the target model to determine the tree species to which each tree belongs.
[0057] The target model includes: multiple diameter at breast height (DBH) identification sub-models corresponding to multiple tree species, one DBH sub-model corresponding to one tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model; the DBH identification sub-model is used to identify the DBH of each tree; the DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-model.
[0058] Understandably, each tree species has a corresponding diameter at breast height (DBH) identification sub-model to calculate the DBH of trees belonging to that species. For example, the DBH identification sub-model for Chinese fir is used to calculate the DBH of Chinese fir, and the DBH identification sub-model for Masson pine is used to calculate the DBH of Masson pine.
[0059] For example, after obtaining the tree height and canopy closure of each tree, these values can be input into the target model to determine the tree species and diameter at breast height (DBH) of each tree. Then, based on the tree height, DBH, and other information of each tree, the timber yield of each tree can be calculated.
[0060] For example, after step 102 above, the single-tree survey method based on lidar point cloud data provided in this application embodiment may further include the following step 103:
[0061] Step 103: Based on the tree species to which each tree belongs, calculate the number of trees corresponding to each tree species in the target area, and calculate the timber yield of each tree based on its height and diameter at breast height.
[0062] For example, after calculating the tree species, tree height, and diameter at breast height (DBH) of each tree, it is possible to base the results on... Figure 2 The yield comparison table shown calculates the yield of each tree.
[0063] Optionally, in the embodiments of this application, the correspondence between tree height and diameter at breast height (DBH) is different for each tree species under different site conditions. Therefore, the DBH normal distribution function corresponding to each DBH identification sub-model needs to be matched with the site conditions of the target area.
[0064] For example, prior to step 102 above, the normal distribution function of chest diameter corresponding to each chest diameter sub-model can also be constructed through steps 104 to 106:
[0065] Step 104: Obtain the site conditions of the target area, and determine the normal distribution constraints corresponding to each diameter at breast height (DBH) identification sub-model based on the site conditions.
[0066] For example, the site conditions of the target area can be obtained by querying relevant meteorological data and other information. Under the same tree age, the diameter at breast height (DBH) value of the same tree species conforms to a normal distribution. However, under different site conditions, the normal distribution of DBH for each tree species is different. Therefore, it is necessary to determine the normal distribution constraints corresponding to each DBH identification sub-model based on the site conditions of the target area and the number of years (i.e., tree age) of the trees planted in the target area.
[0067] Step 105: Based on the normal distribution constraints corresponding to each diameter at breast height (DBH) identification sub-model, randomly generate the expected value and variance of each DBH identification sub-model.
[0068] Step 106: Based on the expected value and variance of each diameter at breast height (DBH) identification sub-model, construct the normal distribution function of DBH for each DBH identification sub-model.
[0069] For example, the above-mentioned normal distribution constraint is used to constrain the expected value and variance of the normal distribution function corresponding to each caliper diameter identification sub-model. After obtaining the normal distribution constraint for each caliper diameter identification sub-model, the expected value and variance of each caliper diameter identification sub-model can be randomly generated within the constraint range based on the constraint. Then, based on the expected value and variance of each caliper diameter identification sub-model, the normal distribution function of caliper diameter corresponding to each caliper diameter identification sub-model can be constructed.
[0070] Optionally, in this embodiment of the application, the diameter at breast height (DBH) of trees can be calculated using the DBH identification sub-model in the target model, and the DBH value calculated by the DBH identification sub-model can be verified using the DBH normal distribution function corresponding to the DBH identification sub-model.
[0071] Specifically, step 102 above may also include the following steps: 102a, 102b and 102c:
[0072] Step 102a: Input the target parameters corresponding to the target trees into the diameter at breast height (DBH) identification sub-model corresponding to each tree species to obtain multiple candidate DBH values.
[0073] One of the chest diameter recognition sub-models outputs a candidate chest diameter value.
[0074] For example, for the tree species identification calculated for any target tree in the target area mentioned above, the target parameters of the target tree can be input into each diameter at breast height (DBH) identification sub-model to obtain multiple candidate DBH values.
[0075] Specifically, the diameter at breast height (DBH) identification sub-model for each tree species is constructed based on the following formula:
[0076] DBH = aH a1 cc a2 S a3 N a4 (Formula 1)
[0077] Where DBH is diameter at breast height (DBH), H is tree height, cc is canopy closure, S is site conditions, N is tree age obtained based on age group, and a, a1, a2, a3, and a4 are coefficients in the coefficient group, with different coefficient groups corresponding to different tree species.
[0078] For example, Formula 1 above is used to characterize the correspondence between diameter at breast height (DBH) and tree height, canopy closure, site conditions, and tree age.
[0079] Step 102b: Based on the normal distribution function of the thorax diameter corresponding to each thorax diameter identification sub-model, calculate the target difference between the candidate thorax diameter value corresponding to each thorax diameter identification sub-model and the median of the normal distribution function corresponding to each thorax diameter identification sub-model.
[0080] Step 102c: Select the candidate diameter at breast height (DBH) value with the smallest target difference from the multiple candidate DBH values and determine the DBH identification sub-model corresponding to it as the target DBH identification sub-model, and determine the tree species corresponding to the target DBH identification sub-model as the tree species to which the target forest belongs.
[0081] For example, after obtaining the candidate diameter at breast height (DBH) values output by each DBH identification sub-model, it is also necessary to perform a rationality verification on the candidate DBH values output by each DBH identification sub-model in order to determine the tree species to which the target forest tree belongs.
[0082] The single-tree survey method based on lidar point cloud data provided in this application first acquires lidar point cloud data of trees within a target area, and determines target parameters for each tree within the target area based on the lidar point cloud data. The target parameters include tree height, canopy closure, and age group. Then, the target parameters for each tree are input into a target model to determine the tree species to which each tree belongs. The target model includes multiple diameter-at-breast-width (DBH) identification sub-models corresponding to multiple tree species, with one DBH sub-model per tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model. The DBH identification sub-models are used to identify the DBH of each tree; the DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-models. In this way, the tree species can be accurately identified through the target model, greatly improving the accuracy of tree species identification based on lidar data collected by UAVs.
[0083] It should be noted that the single-tree survey method based on lidar point cloud data provided in this application embodiment can be executed by a single-tree survey device based on lidar point cloud data, or by a control module within that single-tree survey device for executing the single-tree survey method based on lidar point cloud data. This application embodiment uses the execution of the single-tree survey method based on lidar point cloud data by a single-tree survey device based on lidar point cloud data as an example to illustrate the single-tree survey device based on lidar point cloud data provided in this application embodiment.
[0084] It should be noted that, in the embodiments of this application, the single-tree survey methods based on lidar point cloud data shown in the accompanying drawings are all illustrated by way of example with reference to one accompanying drawing in the embodiments of this application. In specific implementation, the single-tree survey methods based on lidar point cloud data shown in the accompanying drawings of the above methods can also be implemented in conjunction with any other accompanying drawings that can be combined with the above embodiments, which will not be elaborated here.
[0085] The single-tree survey device based on lidar point cloud data provided in this application is described below. The single-tree survey method based on lidar point cloud data described below can be referred to in correspondence with the single-tree survey method based on lidar point cloud data described above.
[0086] Figure 3 A schematic diagram of the structure of the single-tree survey device based on lidar point cloud data provided in the embodiments of this application is shown below. Figure 3 As shown, it specifically includes:
[0087] The acquisition module 301 is used to acquire radar point cloud data of trees within the target area and determine the target parameters of each tree within the target area based on the radar point cloud data; the target parameters include: tree height, canopy closure, and age group; the calculation module 302 is used to input the target parameters of each tree into the target model to determine the tree species to which each tree belongs; wherein, the target model includes: multiple diameter at breast height (DBH) identification sub-models corresponding to multiple tree species, one DBH sub-model corresponding to one tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model; the DBH identification sub-model is used to identify the DBH of each tree; the DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-model.
[0088] Optionally, the acquisition module 301 is specifically used to analyze the radar point cloud data, analyze the elevation value of each point and the distance between each point and other points through a point cloud segmentation algorithm, and determine multiple sub-point clouds based on the elevation value of each point and the distance between each point and other points; each sub-point cloud corresponds to a tree; the acquisition module 301 is also specifically used to analyze each of the multiple sub-point clouds, determine the tree height and canopy closure of the tree corresponding to each sub-point cloud, and determine the age group of each tree in the target area by querying afforestation data.
[0089] Optionally, the device further includes: a construction module; the acquisition module 301 is further configured to acquire the site conditions of the target area and determine the normal distribution constraints corresponding to each diameter at breast height (DBH) identification sub-model based on the site conditions; the construction module is configured to randomly generate the expected value and variance of each DBH identification sub-model based on the normal distribution constraints corresponding to each DBH identification sub-model; the construction module is further configured to construct the normal distribution function of DBH corresponding to each DBH identification sub-model based on the expected value and variance of each DBH identification sub-model.
[0090] Optionally, the calculation module 302 is specifically used to input the target parameters corresponding to the target forest tree into the diameter at breast height (DBH) identification sub-model corresponding to each tree species to obtain multiple candidate DBH values; each DBH identification sub-model outputs a candidate DBH value; the calculation module 302 is also specifically used to calculate the target difference between the candidate DBH value corresponding to each DBH identification sub-model and the median of the normal distribution function corresponding to each DBH identification sub-model based on the normal distribution function of DBH corresponding to each DBH identification sub-model; the calculation module 302 is also specifically used to select the DBH identification sub-model corresponding to the candidate DBH value with the smallest target difference from the multiple candidate DBH values and determine it as the target DBH identification sub-model, and determine the tree species corresponding to the target DBH identification sub-model as the tree species to which the target forest tree belongs.
[0091] Optionally, the diameter at breast height (DBH) identification sub-model for each tree species is constructed based on the following formula:
[0092] DBH = aH a1 cc a2 S a3 N a4 (Formula 1)
[0093] Where DBH is diameter at breast height (DBH), H is tree height, cc is canopy closure, S is site conditions, N is tree age obtained based on age group, and a, a1, a2, a3, and a4 are coefficients in the coefficient group, with different coefficient groups corresponding to different tree species.
[0094] Optionally, the calculation module 302 is further configured to calculate the number of trees corresponding to each tree species in the target area based on the tree species to which each tree belongs, and to calculate the timber yield of each tree based on the tree height and diameter at breast height of each tree.
[0095] The single-tree survey device based on lidar point cloud data provided in this application first acquires lidar point cloud data of trees within a target area, and determines the target parameters of each tree within the target area based on the lidar point cloud data. The target parameters include tree height, canopy closure, and age group. Then, the target parameters of each tree are input into a target model to determine the tree species to which each tree belongs. The target model includes multiple diameter-at-breast-width (DBH) identification sub-models corresponding to multiple tree species, with one DBH sub-model corresponding to one tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model. The DBH identification sub-model is used to identify the DBH of each tree; the DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-model. In this way, the tree species can be accurately identified through the target model, greatly improving the accuracy of tree species identification based on lidar data collected by UAVs.
[0096] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other through the communication bus 440. The processor 410 can call logic instructions in the memory 430 to execute a single-tree survey method based on lidar point cloud data. This method includes: acquiring lidar point cloud data of trees within a target area, and determining target parameters for each tree within the target area based on the lidar point cloud data; the target parameters include: tree height, canopy closure, and age group; inputting the target parameters of each tree into a target model to determine the tree species to which each tree belongs; wherein the target model includes: multiple diameter-at-breast-width (DBH) identification sub-models corresponding to multiple tree species, one DBH sub-model corresponding to one tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model; the DBH identification sub-model is used to identify the DBH of each tree; the DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-model.
[0097] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0098] On the other hand, this application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by the computer, the computer can execute the single-tree survey method based on lidar point cloud data provided by the above methods. The method includes: acquiring lidar point cloud data of trees in a target area, and determining target parameters for each tree in the target area based on the lidar point cloud data; the target parameters include: tree height, canopy closure, and age group; inputting the target parameters of each tree into a target model to determine the tree species to which each tree belongs; wherein, the target model includes: multiple diameter at breast height (DBH) identification sub-models corresponding to multiple tree species, one DBH sub-model corresponding to one tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model; the DBH identification sub-model is used to identify the DBH of each tree; the DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-model.
[0099] Furthermore, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the aforementioned methods for single-tree surveys based on lidar point cloud data. The method includes: acquiring lidar point cloud data of trees within a target area, and determining target parameters for each tree within the target area based on the lidar point cloud data; the target parameters include: tree height, canopy closure, and age group; inputting the target parameters of each tree into a target model to determine the tree species to which each tree belongs; wherein the target model includes: multiple diameter-at-breast-width (DBH) identification sub-models corresponding to multiple tree species, one DBH sub-model corresponding to one tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model; the DBH identification sub-model is used to identify the DBH of each tree; the DBH normal distribution function is used to verify the reasonableness of the DBH identified by the DBH identification sub-model.
[0100] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0101] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. 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. Such 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.
Claims
1. A method for single tree surveying based on lidar point cloud data, characterized in that, include: Acquire radar point cloud data of trees within the target area, and determine the target parameters of each tree within the target area based on the radar point cloud data; The target parameters include: tree height, canopy closure, and age group; Input the target parameters of each tree into the target model to determine the tree species to which each tree belongs; The target model includes: multiple diameter at breast height (DBH) identification sub-models corresponding to multiple tree species, one DBH sub-model corresponding to one tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model; the DBH identification sub-model is used to identify the DBH of each tree; the DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-model. The step of inputting the target parameters of each tree into the target model to determine the tree species to which each tree belongs includes: The target parameters corresponding to the target trees are input into the diameter at breast height (DBH) identification sub-model corresponding to each tree species to obtain multiple candidate DBH values; each DBH identification sub-model outputs one candidate DBH value. Based on the normal distribution function of chest diameter corresponding to each chest diameter recognition sub-model, calculate the target difference between the candidate chest diameter value corresponding to each chest diameter recognition sub-model and the median of the normal distribution function corresponding to each chest diameter recognition sub-model; The candidate diameter at breast height (DBH) value with the smallest target difference is selected from the multiple candidate DBH values and the corresponding DBH identification sub-model is determined as the target DBH identification sub-model. The tree species corresponding to the target DBH identification sub-model is determined as the tree species to which the target forest belongs.
2. The method of claim 1, wherein, The determination of the tree height and canopy density of each tree within the target area based on the radar point cloud data includes: The radar point cloud data is analyzed by using a point cloud segmentation algorithm to analyze the elevation value of each point and its distance from other points, and multiple sub-point clouds are determined based on the elevation value of each point and its distance from other points; each sub-point cloud corresponds to a tree. Each of the multiple sub-point clouds is analyzed to determine the tree height and canopy closure of the corresponding trees in each sub-point cloud, and the age group of each tree in the target area is determined by querying afforestation data.
3. The method of claim 1, wherein, Before inputting the target parameters of each tree into the target model to determine the tree species to which each tree belongs, the method further includes: Obtain the site conditions of the target area, and determine the normal distribution constraints corresponding to each diameter at breast height (DBH) identification sub-model based on the site conditions; Based on the normal distribution constraints corresponding to each diameter at breast height (DBH) identification sub-model, the expected value and variance of each DBH identification sub-model are randomly generated. Based on the expected value and variance of each diameter at breast height (DBH) identification sub-model, a normal distribution function for DBH corresponding to each DBH identification sub-model is constructed.
4. The method of claim 1, wherein, The target parameters corresponding to the target trees are input into the diameter at breast height (DBH) identification sub-model corresponding to each tree species to obtain multiple candidate DBH values, including: The diameter at breast height (DBH) identification sub-model for each tree species is constructed based on the following formula: (Equation One) wherein, is the breast height diameter, H is the tree height, cc is the canopy density, S is the site condition, N is the tree age based on the age group, a , , , , are coefficients in the coefficient array, and different tree species correspond to different coefficient arrays.
5. The method of claim 1, wherein, After inputting the target parameters of each tree into the target model and determining the tree species to which each tree belongs, the method further includes: Based on the tree species to which each tree belongs, the number of trees corresponding to each tree species in the target area is calculated, and the timber yield of each tree is calculated based on the tree height and diameter at breast height of each tree.
6. A single-tree survey device based on lidar point cloud data, characterized in that, The device includes: The acquisition module is used to acquire radar point cloud data of trees within a target area, and determine the target parameters of each tree within the target area based on the radar point cloud data; the target parameters include: tree height, canopy closure and age group; The calculation module is used to input the target parameters of each tree into the target model to determine the tree species to which each tree belongs; The target model includes: multiple diameter at breast height (DBH) identification sub-models corresponding to multiple tree species, one DBH sub-model corresponding to one tree species, and a DBH normal distribution function corresponding to each DBH identification sub-model; the DBH identification sub-model is used to identify the DBH of each tree; the DBH normal distribution function is used to verify the rationality of the DBH identified by the DBH identification sub-model. The calculation module is specifically used to input the target parameters corresponding to the target trees into the diameter at breast height (DBH) identification sub-model corresponding to each tree species to obtain multiple candidate DBH values; each DBH identification sub-model outputs a candidate DBH value. The calculation module is further configured to calculate the target difference between the candidate chest diameter value corresponding to each chest diameter identification sub-model and the median of the normal distribution function corresponding to each chest diameter identification sub-model, based on the chest diameter normal distribution function corresponding to each chest diameter identification sub-model. The calculation module is further configured to select the candidate diameter at breast height (DBH) value with the smallest target difference from the plurality of candidate DBH values, determine the DBH identification sub-model corresponding to the candidate DBH value as the target DBH identification sub-model, and determine the tree species corresponding to the target DBH identification sub-model as the tree species to which the target forest belongs.
7. The apparatus according to claim 6, characterized in that, The device further includes: a construction module; The acquisition module is also used to acquire the site conditions of the target area and determine the normal distribution constraint conditions corresponding to each diameter at breast height (DBH) identification sub-model based on the site conditions. The construction module is used to randomly generate the expected value and variance of each caliper diameter recognition sub-model based on the normal distribution constraint conditions corresponding to each caliper diameter recognition sub-model. The construction module is also used to construct the normal distribution function of chest diameter for each chest diameter recognition sub-model based on the expected value and variance of each chest diameter recognition sub-model.
8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the single-tree survey method based on lidar point cloud data as described in any one of claims 1 to 5.
9. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the steps of the single-tree survey method based on lidar point cloud data as described in any one of claims 1 to 5.