Three-dimensional Reconstruction Of Standing Tree And Attributes Measurement Method

The diameter at breast height(DBH),height and crown width of standing tree are important measurement factors in forest resource inventory and an important basis for evaluating the growth of trees in sampling sites.The current standing tree factor measurement methods have disadvantages such as low degree of automation and subjective influence on the measurement results,while the standing tree factor measurement equipment based on three-dimensional reconstruction technology is expensive and difficult to popularize.Therefore,this thesis proposes a three-dimensional reconstruction and factor measurement method for standing trees.This method only uses a smartphone to shoot a video around a scene containing multiple standing trees,obtains its keyframe sequence images,uses the feature point matching relationship between the video keyframes,and performs three-dimensional reconstruction of multiple standing trees based on structure from motion algorithm(Sf M),and finally based on the 3D point cloud of multiple standing trees,realizes automatic measurement of the DBH,height,and crown width of multiple standing trees.The main research contents of this thesis includes(1)Research on the method of feature point matching of standing tree images.Enhance the image of the trunk of the standing tree to obtain more detailed information,then use RGBSIFT algorithm for feature points detection and description,and finally complete the feature point matching of the standing tree based on the geometric position constraint,thereby improving the ability to correctly match the feature points of standing tree images,so as to improve the stability of the camera parameter estimation during the subsequent three-dimensional reconstruction of the standing tree.(2)Research on the three-dimensional reconstruction method of multiple standing trees based on structure from motion.Automatically extract the key frame image sequence according to the similarity calculation between the standing wood video frames,and complete the camera parameter estimation and the sparse 3D point cloud of the generated tree through the structure from motion algorithm,and the sparse 3D point cloud is densified and down-sampling,denoising,coordinate correction,and segmentation to obtain a 3D point cloud model of multiple standing trees that is finally used for standing tree factor measurement.(3)Research on the measurement method of multi standing tree factors based on 3D point cloud.Using the standing tree 3D point cloud model to extract the key points of the standing tree,the spatial structure relationship between the key 3D points of the standing tree is deduced,so as to calculate the DBH,tree height and crown width of multiple standing trees to realize the efficient and convenient measurement of the standing tree factor.The method in this thesis is applied to the measurement of DBH,tree height,and crown width of standing trees and compared with their actual values.The results show that the average relative errors of DBH,height and crown width are 2.66%,1.85% and 3.99%,respectively,and its measurement accuracy can meet the requirements of the forest resource planning and design survey(the second type of survey).