CN-122024043-A - Vegetation growth evaluation method and system based on laser point cloud identification

CN122024043ACN 122024043 ACN122024043 ACN 122024043ACN-122024043-A

Abstract

The invention discloses a vegetation growth evaluation method and a vegetation growth evaluation system based on laser point cloud identification, wherein the method comprises the following operation steps of collecting two-period point cloud data of an interval time period for a target area; the two-period point cloud data comprise first-period point cloud data and second-period point cloud data, comprehensive difference values between point cloud densities of the two-period point cloud data are determined, a shielding candidate region is determined, space statistical features are analyzed on the shielding candidate region, vegetation category point cloud analysis growth parameter sets are divided by the aid of the space statistical features, plant growth simulation is conducted on the growth parameter sets, and shielding results of vegetation point clouds after growth simulation are judged. And screening the shielding candidate area through the point cloud to obtain a vegetation point cloud, and reflecting the shielding relation between the crown and the shrub by utilizing the vegetation point cloud so as to determine whether shielding exists between the crown and the shrub.

Inventors

ZHAO KUNYU
ZHANG YUNPENG
LI BU
HUANG XIANGYI
LIU MENGYI
Wu yaxing
CHEN QIHONG
LIU WEIYU
CHEN SHIYUN
FENG HAO

Assignees

重庆市送变电工程有限公司

Dates

Publication Date: 20260512
Application Date: 20260114

Claims (10)

1. The vegetation growth evaluation method based on laser point cloud identification is characterized by comprising the following operation steps: Acquiring two-period point cloud data of an interval time period from a target area, wherein the two-period point cloud data comprise first-period point cloud data and second-period point cloud data; Determining an occlusion candidate region according to the comprehensive difference value between the point cloud densities of the point cloud data of the two periods, analyzing space statistical characteristics of the occlusion candidate region, dividing the occlusion candidate region into vegetation category point cloud analysis growth parameter sets according to the space statistical characteristics, performing plant growth simulation on the growth parameter sets to obtain an occlusion result of the vegetation point cloud after the growth simulation.
2. The vegetation growth evaluation method based on laser point cloud identification according to claim 1, wherein for the comprehensive difference value between the point cloud densities of the two-period point cloud data, an occlusion candidate area is determined, and the specific operation steps are as follows: The method comprises the steps of setting voxels of a multi-scale grid for two-period point cloud data respectively, calculating according to the number of point clouds of each voxel in the two-period point cloud data to obtain a first-period point cloud density and a second-period point cloud density respectively, calculating a comprehensive difference value by utilizing the first-period point cloud density and the second-period point cloud density, and screening the voxels according to the comprehensive difference value of each voxel and a preset difference density threshold to serve as shielding candidate areas.
3. The vegetation growth evaluation method based on laser point cloud identification according to claim 2, wherein the analysis of the spatial statistical features of the occlusion candidate area comprises the following steps: Determining coordinates of voxels in each shielding candidate region, searching maximum and minimum values of the coordinates of the voxels in each shielding candidate region, and obtaining a boundary frame of the shielding candidate region; Calculating the number of the point clouds in each shielding candidate area to be used as the area point cloud density; and the regional area boundary descriptor and the regional point cloud density combination are used as spatial statistical characteristics.
4. The vegetation growth evaluation method based on laser point cloud identification according to claim 3, wherein the method comprises the steps of dividing the vegetation category point cloud analysis growth parameter set for the shielding candidate area by using the space statistical characteristics, performing plant growth simulation for the growth parameter set, and judging the shielding result of the vegetation point cloud after the growth simulation, wherein the specific operation steps are as follows: the method comprises the steps of expanding the shielding candidate region by utilizing the space statistical characteristics to obtain a three-dimensional round buffer region, analyzing the round buffer region to obtain multi-dimensional characteristics, inputting the multi-dimensional characteristics into a point cloud segmentation model constructed in advance to divide vegetation category point clouds, dividing the round buffer region into space scale regions, analyzing a growth parameter set of the vegetation category point clouds in each space scale region, carrying out plant growth simulation by utilizing the growth parameter set to update the point cloud segmentation model to obtain vegetation point clouds again, and carrying out clustering calculation on the vegetation point clouds to judge a shielding result.
5. The vegetation growth evaluation method based on laser point cloud identification according to claim 4, wherein the coverage candidate region is expanded by using the spatial statistical feature to obtain a three-dimensional sphere buffer region, the sphere buffer region is analyzed to obtain a multidimensional feature, and the specific operation steps are as follows: calculating the shielding candidate region by using the space statistical characteristics to obtain a buffer radius; determining three-dimensional coordinates of the point clouds in the shielding candidate area, and judging whether the three-dimensional coordinates of all the point clouds are located in the round body buffer area or not; if so, calculating the density of the buffer point cloud for the point cloud sitting in the round buffer area; acquiring a laser echo intensity value from laser point cloud equipment; the round buffer area corresponds to the multispectral image in pixel point coordinates, and color attributes of corresponding pixel points are extracted; And forming the buffer point cloud density, the laser echo intensity value and the color attribute to form the multidimensional characteristic of the round buffer area.
6. The vegetation growth evaluation method based on laser point cloud identification according to claim 5, wherein the multidimensional features are input into a pre-built point cloud segmentation model to divide vegetation category point clouds, the circle buffer area is divided into space scale intervals, and a growth parameter set of the vegetation category point clouds in each space scale interval is analyzed, and the specific operation steps are as follows: The method comprises the steps of utilizing a deep learning network as a frame to construct a point cloud segmentation model in advance, inputting multidimensional features of a round buffer area into the point cloud segmentation model, and outputting vegetation class point clouds in the round buffer area; Comparing the vegetation category point cloud height increasing rate in the same space scale interval in the first period point cloud data and the second period point cloud data, projecting the vegetation category point cloud to a horizontal plane, and acquiring a canopy expansion direction by using the vegetation category point cloud of the projection horizontal plane; Combining the height growth rate and the canopy expansion direction to form a growth trend parameter; And forming a growth parameter set by using the growth trend parameters and the multi-scale feature vectors.
7. The vegetation growth evaluation method based on laser point cloud identification according to claim 6, wherein the plant growth simulation is performed by using the growth parameter set to update the point cloud segmentation model, and vegetation point clouds are obtained again: The method comprises the steps of simulating plant growth by using a growth parameter set to construct a branch and leaf geometric model, supplementing point clouds to the branch and leaf geometric model, updating a point cloud segmentation model by using the supplementing point clouds, analyzing multidimensional features of the round buffer area again to obtain vegetation point clouds, clustering the vegetation point clouds to obtain independent point cloud clusters, calculating the coordinate position of the height of each independent point cloud cluster, and judging a shielding result according to the coordinate position of the height.
8. The vegetation growth evaluation method based on laser point cloud identification of claim 7, wherein the plant growth simulation is performed by using the growth parameter set to construct a branch and leaf geometric model, and the branch and leaf geometric model is supplemented with point cloud, and the specific operation steps are as follows: setting L system rules by using the growth parameter set to introduce a sunward mechanism; collecting peripheral point clouds of a target area, and extracting multidimensional features from the peripheral point clouds; Correcting the growth parameter set by utilizing the error of the multi-dimensional characteristic, and constructing a branch-leaf geometric model by utilizing the corrected growth parameter set and the branch-leaf geometric structure; Setting sampling density for the branch-leaf geometric model, randomly sampling the position points with highest sampling point density for the branch-leaf geometric model according to the sampling density, and determining the current sampling coordinates; Calculating the point cloud density of the peripheral point cloud and the complementary point cloud respectively, and further calculating a point cloud density difference value; randomly sampling the branch-leaf geometric model again by using the new sampling density, and calculating a new point cloud density difference value; if yes, judging that the new complementary point cloud and the branch and leaf geometric model are completely and seamlessly connected, and obtaining the final complementary point cloud.
9. The vegetation growth evaluation method based on laser point cloud identification according to claim 8, wherein the method is characterized by updating a point cloud segmentation model by using supplementary point clouds, analyzing the multidimensional features of the round buffer area again to obtain vegetation point clouds, clustering the vegetation point clouds to obtain independent point cloud clusters, calculating the coordinate position of the height of each independent point cloud cluster, and judging the shielding result according to the height coordinate position, wherein the method comprises the following specific operation steps: correcting and updating the point cloud segmentation model by utilizing the final supplementary point cloud, and re-inputting the multidimensional features of the round buffer area into the updated point cloud segmentation model to perform point cloud segmentation to obtain final vegetation point clouds; The method comprises the steps of determining the coordinate position of a vertex point cloud in each independent point cloud cluster as the coordinate position of the independent point cloud cluster, sorting the heights of the coordinate positions of all the independent point cloud clusters, calculating a size percentile, taking the size percentile of the independent point cloud cluster with the lowest height in the size percentile as a low percentile, and taking all the other size percentiles as upper-layer percentiles; The method comprises the steps of calculating the area of each independent point cloud cluster of each height, calculating the area overlapping degree by utilizing the area of a low percentile and the area of each other upper-layer size percentile, and determining that the areas of the independent point cloud clusters of the two height size percentiles are shielded from each other according to the area overlapping degree.
10. A vegetation growth evaluation system based on laser point cloud identification is characterized by comprising an acquisition module, an analysis module, a control module and a control module, wherein the acquisition module is used for acquiring vegetation; The acquisition module is used for acquiring two-period point cloud data of an interval period of time for the target area, wherein the two-period point cloud data comprise first-period point cloud data and second-period point cloud data; The analysis module is used for determining an occlusion candidate area for the comprehensive difference value between the point cloud densities of the point cloud data of the two periods, analyzing space statistical characteristics for the occlusion candidate area, dividing vegetation category point cloud analysis growth parameter sets for the occlusion candidate area by utilizing the space statistical characteristics, performing plant growth simulation for the growth parameter sets, and judging occlusion results of the vegetation point clouds after growth simulation.

Description

Vegetation growth evaluation method and system based on laser point cloud identification Technical Field The invention relates to the field of operation and maintenance of power transmission lines, in particular to a vegetation growth assessment method and system based on laser point cloud identification. Background In a crisscross power network, the safe and stable operation of an overhead transmission line is important. While vegetation under the lines and in the hallways is often a potential source of dynamic changes. The control of vegetation growth and species assessment, especially the "line tree relationship" is the first line of defense against grid accidents. When vegetation grows too fast to approach a power transmission line, problems such as circuit conduction faults or forest fire caused by breakdown of trees can be caused. Accurate identification of vegetation, particularly fast-growing tree species and flammable arbor vegetation, is an important detection indicator for vegetation growth assessment. Traditional vegetation growth assessment methods often depend on manual investigation or image processing technology based on optical remote sensing data, but in areas covered by complex terrains or high-density vegetation, the methods are often interfered by external factors such as weather, illumination change, visual angles and the like, so that assessment accuracy and reliability are insufficient. Particularly in environments where multiple vegetation coexist, such as areas where different vegetation types are interwoven, such as trees and shrubs, a great challenge is presented in assessing occlusion effects. In order to overcome the defects of the traditional method, the laser point cloud technology is widely applied to vegetation growth assessment. The laser scanning technology is used for accurately measuring three-dimensional space coordinates of the surface of an object by emitting laser beams from different angles and receiving reflected signals through a laser sensor, so that high-precision point cloud data are constructed. The density and the precision of the point cloud data can reflect the three-dimensional form of vegetation, and can be compared in different time periods to monitor the vegetation growth condition in real time. However, studies have found that tree occlusion occurs during vegetation growth monitoring and assessment, particularly in a symbiotic environment of the arbor and shrub type, where arbor vegetation often occludes a partial area of shrub vegetation, because arbor typically has a significantly upstanding trunk, typically over 6 meters in height, which can be subdivided into small, medium, da Qiao and great trees, such as camphor tree, ginkgo, and the like. The shrubs have no obvious trunks, the plants are short in height and are generally lower than 6 meters. For example, the woods of northeast forests can be clearly classified into two main categories, conifer and broadleaf. The most common arbor is conifer, pinus sylvestris, etc., whereas the most common shrubs are hazelnut, acanthopanax, etc. The above mentioned case of the arbor layer shielding the shrub layer has an important influence on the accurate assessment of the arbor growth and the interaction of the plant with the environment. The traditional evaluation method is difficult to effectively quantify the shielding effect, and the shielding relation between different vegetation types cannot be accurately distinguished. Disclosure of Invention The invention aims to provide a vegetation growth evaluation method and system based on laser point cloud identification, which solve the technical problems pointed out in the prior art. The invention provides a vegetation growth evaluation method based on laser point cloud identification, which comprises the following operation steps: Acquiring two-period point cloud data of an interval time period from a target area, wherein the two-period point cloud data comprise first-period point cloud data and second-period point cloud data; Determining an occlusion candidate region according to the comprehensive difference value between the point cloud densities of the point cloud data of the two periods, analyzing space statistical characteristics of the occlusion candidate region, dividing the occlusion candidate region into vegetation category point cloud analysis growth parameter sets according to the space statistical characteristics, performing plant growth simulation on the growth parameter sets to obtain an occlusion result of the vegetation point cloud after the growth simulation. The invention also provides a vegetation growth evaluation system based on laser point cloud identification, which comprises an acquisition module, an analysis module, a control module and a control module, wherein the acquisition module is used for acquiring vegetation growth information of the vegetation; The acquisition module is used for acquiring two-period point cloud data of an int