3D segmentation of full waveform LiDAR data for single tree detection using normalized cut

The study highlights a novel method to segment single trees in 3Dfrom dense airborne full waveform LIDAR data using the normalized cut segmentation. The key idea is to subdivide the treearea in a voxel space and to setup a bipartite graph which is formed by the voxels and similarity measures between the voxels. Thenormalized cut segmentation divides the graph hierarchically into segments which have a minimum similarity among each other andwhose members (=voxels) have a maximum similarity. The solution is found by solving a corresponding generalized eigenvalueproblem and an appropriate binarization of the solution vector. We applied the method to small-footprint full waveform data that havebeen acquired in the Bavarian Forest National Park with amean point density of 25 points per m² in leaf-off situation. Thesegmentation procedure is evaluated in different steps. First, alinear discriminant analysis shows that the mean intensity of the voxelsderived from the full waveform data contributes significantly tothe segmentation of deciduous and coniferous tree segments. Second, asample-based sensitivity analysis examines the best value of themost important control parameter that stops the division process ofthe graph. Third, we show examples how the segmentation cancope with even difficult situations. We also discuss examples showingthe limits of the current implementation. Finally, we present the detection rate of the new method in controlled tests using referencedata. If we compare the new method to a standard watershed-basedsegmentation approach the overall improvement for all tree layersis 9%. However, the biggest improvement can be achievedin the intermediate layer with 14% and in the lower layer with 16%showing clearly the advantage of the new approach to a 3D segmentation of single trees.     «

The study highlights a novel method to segment single trees in 3Dfrom dense airborne full waveform LIDAR data using the normalized cut segmentation. The key idea is to subdivide the treearea in a voxel space and to setup a bipartite graph which is formed by the voxels and similarity measures between the voxels. Thenormalized cut segmentation divides the graph hierarchically into segments which have a minimum similarity among each other andwhose members (=voxels) have a maximum similarity. The so...     »