Tree measurement and reconstruction method based on single three-dimensional laser scanning

Abstract

The invention provides a method for measuring and reconstructing a tree based on single three-dimensional laser scanning, and relates to the technologies of computer graphs and computer vision. The method comprises steps of data preprocessing, axial estimation, three-dimensional framework computation and three-dimensional model generation. The method only utilizes single scanning data of a laser scanner to obtain branch shape and radius information faithful to the primary tree, carries out main operations in a two-dimensional image space, has simple and effective algorithm and accurate reconstruction result, and has important application values in the computer graphics application fields such as virtual reality, computer game, natural scene simulation, urban landscape design and the like and in the fields such as agriculture and forestry investigation, tree measurement, etc.

Description

technical field [0001] The invention relates to the technical fields of computer graphics and computer vision, and is a tree measurement and reconstruction method based on a single three-dimensional laser scan. The reconstruction results are mainly used in computer graphics such as virtual reality, computer games, natural scene simulation, and urban landscape design. Fields of application and agroforestry surveys are used for tree measurement, etc. Background technique [0002] Trees are one of the most common scenes in nature. In order to enhance the realism of the scene, constructing 3D models of trees has become a must in various application fields of computer graphics, such as virtual reality, computer games, natural scene simulation, urban landscape design, etc. indispensable component. In addition, reconstructing the 3D model of branches also plays an important role in agricultural and forestry applications, such as tree parameter measurement, plant growth simulation,...

AI Technical Summary

Problems solved by technology

Due to the inevitable occlusion between the leaves and branches of the tree itself, and the occlusion of other trees or buildings in the surrounding environment, it is difficult to obtain complete tree surface data; in addition, due to the complex geometry and topology of the tree, 3D reconstruction of real trees is challenging

At present, there are not many works in this area, and a common feature of the methods used is that from the perspective of graphics, the model generated for trees with leaves has a better overall visual effect, but due to the occlusion of leaves, most branches The trunk is invisible and has no input information when measuring. The model of this part of the branch, including the shape and diameter of the branch, can only be generated virtually based on botanical knowledge, which is quite different from the original tree shape, and finally a specious model of the original tree is obtained. 3D model, does not have the fidelity of shape and structure

In addition, the skeleton of the visible part of the bottom of the tree is obtained by the seed point distance clustering method, which has poor smoothness and often requires manual adjustment

Images