Method for automatically reconstructing tree structure from ground-based laser radar point cloud

Abstract

The invention provides a method for automatically reconstructing a tree branch structure from a ground-based laser radar point cloud, which is characterized in that in order to accurately fit a tree branch geometric structure and a topological relation from a TLS point cloud, a single tree geometric structure and topology are reconstructed on the basis of a graph theory method in combination with the advantages of prior hypothesis modeling and lightweight expression; the method for reconstructing the fine geometric structure of the individual tree branches and trunks is high in universality and precision. Firstly, triangulation of an input point cloud is completed based on three-dimensional Delaunay, and a shortest path of a directed weighted graph is calculated based on a Dijkstra algorithm. And then extracting an initial tree skeleton by using a minimum spanning tree (MST) algorithm framework, designing a merging algorithm of redundant vertexes and edges, and completing simplification and optimization of the initial skeleton. And finally, based on the tree skeleton, single tree branch structure cylinder fitting and an optimization algorithm thereof are completed, and a fine single tree branch geometric structure is reconstructed.

Description

1. Technical field [0001] The invention relates to a tree reconstruction method, mainly a method for automatically reconstructing a tree branch structure from a ground-based laser radar point cloud. [0002] 2. Technical Background [0003] With the rise of LiDAR software and hardware technology and the gradual reduction in cost, it has been continuously studied and applied in the field of forest ecology. LiDAR point cloud can describe detailed tree geometry information, which provides a data basis for tree modeling with high precision, strong sense of reality and multi-level detail expression. In recent years, some researchers at home and abroad have tried and explored the theory, method and technology of tree modeling based on lidar point cloud, and achieved results that have more forestry advantages than traditional tree modeling methods. At present, it mainly includes clustering thought modeling, graph theory modeling, prior hypothesis modeling, Laplacian modeling and li...

AI Technical Summary

Problems solved by technology

Due to the sensitivity to point cloud noise or point cloud integrity, the loss of canopy information caused by dimensionality reduction operations will lead to wrong branch topological connections or incomplete structures.

Images