Novel Approach for Tree Detection in Japanese Plantation Forests using Drone LiDAR Data

This paper presents a novel tree detection method for plantation forests using drone LiDAR data. By focusing on the understory structure and utilizing point clouds containing tree trunks, we apply hierarchical clustering and principal component analysis algorithms to accurately identify and retain trunk features. Experimental results in dense cypress and cedar plantation forests show improved detection accuracy compared to canopy-based methods in high tree density plots.