Combined Tree Segmentation and Stem Detection Using Full Waveform Lidar Data

The study highlights a new method for the delineation of tree crowns and the detection of stem positions of single trees from dense airborne LIDAR data. At first, we combine a method for surface reconstruction, which robustly interpolates the canopy height model (CHM) from the LIDAR data, with a watershed algorithm. Stem positions of the tallest trees in the tree segments are subsequently derived from the local maxima of the CHM. Additional stem positions in the segments are detected in a 3-step algorithm. First, all the points between the ground and the crown base height are separated. Second, possible stem points are found by hierarchically clustering these points. Third, the stem is reconstructed with a robust RANSAC-based adjustment of the stem points. The method was applied to small-footprint full waveform data, which have a point density of 25 points per m. The detection rate for coniferous trees is 61 % and for deciduous trees 44 %, respectively. 7 % of the detected trees are false positives. The mean positioning error is 0.92 cm, whereas the additional stem detection improves the tree position on average by 22 cm. The analysis of waveform data in the tree structure shows that the intensity and pulse width discriminate stem points, crown points and ground points significantly. Moreover, the mean intensity of stem points turned out to be the most salient feature for the discrimination of coniferous and deciduous trees. * Corresponding author.