| Backpack Laser Scanning(Backpack Li DAR)is a new Laser radar Scanning technology.Due to the continuous development of its backpack-type operation mode and related technical methods,it can play a great role in the investigation of forestry resources,especially in the extraction of single wood DBH.Compared with the airborne laser scanning technology,it can obtain the understory point cloud data more completely,quickly and conveniently.Traditional extraction method of diameter at breast height is by capturing single tree DBH circle fitting point cloud,but in the city park,in order to ensure that transplanting the growth of trees can add outfit support bar for trees,if using the traditional method for DBH extraction produces large error,this article from the problems of the backpack Li DAR point cloud data extraction single tree diameter at breast height,tree higher parameters are studied,The main research contents are summarized as follows:(1)Three different algorithms for single wood segmentation are studied and compared.In this study,on the basis of obtaining single wood point cloud data,single wood segmentation based on regional growth,single wood segmentation based on Euclidean clustering and single wood segmentation based on DBSCAN density clustering were conducted on two groups of point clouds with different data sizes and complexity.The single wood segmentation accuracy and efficiency of the three methods were compared.The results show that the segmentation efficiency based on region growth and Euclidian clustering is better than that based on DBSCAN density clustering,but the segmentation accuracy is inferior to DBSCAN density clustering in the case of overlapping tree point clouds.(2)The extraction method of single tree height was improved.The traditional method of tree height extraction is to find the difference between the maximum and minimum z value of single point cloud.It is found in this study that there are some errors in the tree root point cloud filtering during ground point filtering.If the maximum z value is directly taken as the height of a single tree,the extraction result may be too high because the point is a noise point.In this study,after extracting the maximum z-value point,the point was taken as the waiting point of tree height,and the point density in the neighborhood of the point was added as the judging condition,which could improve the accuracy of tree height extraction.(3)The method of tree DBH extraction in city park was improved.Most trees in urban parks have support rods which greatly affect the accuracy of DBH extraction.In this study,after denoising before DBH extraction,more accurate DBH parameters can be obtained by directly circular fitting of point cloud slices at DBH of trees in natural forests.Aiming at the problem of supporting rod point cloud noise in point cloud slices at DBH of single trees in urban parks,with the help of C++ and Python programming language,combined with PCL and Open 3D library,the improved circle fitting method based on least squares was realized.The DBH point cloud of single wood was determined and then the DBH fitting was carried out by clustering point cloud of slice,calculating the centroid of the cluster and comparing the distance between the centroid of each cluster and other clusters,which greatly improved the DBH extraction result.Combined with the actual project of single tree parameter extraction in Chongqing Guangyang Island City Park,the research results and development program of this paper are practiced.The results show that the research results of this paper can greatly improve the extraction accuracy of single wood parameters in urban parks,and the extraction results meet the accuracy and efficiency requirements of the attached projects,which can be popularized and applied.This paper has 49 sets of figures,14 tables and 59 references. |
PDF Full Text Request