Pipeline Approach to Segmenting Individual Trees from Terrestrial LiDAR Forest Data

LiDAR technology has revolutionized the analysis of forest structure by offering three-dimensional insights essential for effective forest management. This study proposes an innovative pipeline-based tree separation algorithm, specifically designed for autonomous segmentation of individual trees in terrestrial LiDAR-scanned forests. The algorithm encompasses a streamlined, eight-phase process encompassing duplicate and outlier point removal, ground flattening, trunk-crown segmentation, individual tree trunk separation, voxel downsampling, tree separation, and calculation of tree properties. This approach ensures a seamless transition between phases, with each phase's output feeding directly into the next. Our empirical study, employing real-life terrestrial LiDAR datasets, demonstrates the algorithm's effectiveness. This advancement not only elevates digital forest management tools but also has substantial implications for ecological and environmental research.