Application of small-footprint airborne LiDAR data to segment under-story trees in deciduous forests

Airborne LiDAR point cloud representing a forest contains 3D data, from which vertical stand structure can be derived. This paper presents a tree segmentation approach for multi-story stands that iteratively strips canopy layers off the point cloud and segments individual tree crowns within each layer using a digital surface model based tree segmentation method as a building block. We analyze the vertical distributions of LiDAR points within overlapping locales in order to determine the local height thresholds for stripping a canopy layer. Unlike the previous work that stripped stiff layers within constrained areas, the local layering method strips flexible (in thickness and height) canopy layers within unconstrained areas, which can also be utilized as a robust vertical stratification of canopy, independent of the tree segmentation method applied to each layer. Statistical analyses showed that layering strongly improves detecting under-story trees at the cost of moderately increasing over-segmentation rate of the detected under-story trees, while only slightly affecting the segmentation quality of over-story trees. Results obtained from layering the canopy suggest that acquiring denser LiDAR point clouds (becoming affordable due to advancements of the sensor technology and platforms) would allow segmenting under-story trees as accurately as over-story trees. Keywords: LiDAR remote sensing, multi-layered stand, canopy layering, vertical stratification, individual tree segmentation.