TreeLearn

TreeLearn segments individual trees from terrestrial and mobile laser scans, where the scanner sits beneath the canopy rather than above it. It predicts an offset vector for each point toward its tree’s center, then clusters, an approach suited to the dense and occluded geometry a ground-based scan produces.

Ground-based scanning resolves stems and branch structure not visible to airborne lidar. TreeLearn is the counterpart to SegmentAnyTree rather than a competitor.