期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2010
卷号：XXXVIII - Part 8
页码：675-678
出版社：Copernicus Publications
摘要：There is an increasing need for an accurate forest inventory for assessment of forest condition. Due to such a demand, an airborne LiDAR instruments have become increasingly popular in forest inventory. Tree detection and tree crown delineation from an airborne LiDAR has mostly focusing on utilizing the rasterized canopy height model. This paper presents a method for individual tree crown delineation based on triangulated irregular network (TIN) derived from point cloud. Tree crown borders were estimated by abrupt changes of surface normal vector calculated from the estimated tree top position. The vector-based approach was able to avoid trial and error such as a definition of mesh size or an interpolation method. The developed algorithm was tested in a park forest which includes various tree crown size. The results indicate that the developed algorithm was able to avoid the effect of holes in the tree crown area and was able to delineate the isolated tree crown regardless of tree crown size. However, it is necessary to improve the accuracy of estimation of tree top position and the decision method of search area
关键词：Park; Crown size; hole; Vector-based algorithm; TIN; 3D delineation
