A new method for individual tree measurement from airborne LiDAR

This paper presents a new method for individual tree measurement from Airborne LiDAR data. This method involves 3 steps; 1) individual tree crown delineation based on density of high points (DHP), 2) tree filtering, and 3) measurement of tree trunk diameter at breast height (DBH). In the second step, a special tree filtering algorithm is introduced which combines a histogram analysis and region growing (RG) segmentation method. In forest area, undergrowth vegetation is considered as noise and it should be removed to ease the DBH measurement process of trees. The DBH measurement on point cloud is done based on two steps; 1) three-dimensional line fitted on points of tree trunk, and 2) histogram analysis of distances between points and the line. It shows that more than 60% trees are successfully filtered and compared to the actual DBH measurement in the field the DBH estimations on point cloud have the root mean square error of 0.18 m.