评价指标结果表明，点云距离阈值分割方法在落叶松的单木分割结果中表现最优

Evaluation results show that, from the point cloud performance threshold segmentation results in a single timber divided Larch optimal (average F = 0.94), where young larch multiscale segmentation accuracy cloud point method quite a distance threshold. Larch relatively regular crown shape, the gap is relatively clear neighboring trees, and more Individual Tree high average, high SFM feature point matching method accuracy, so set the appropriate distance threshold point cloud can achieve a better segmentation results.

The evaluation index results show that the point cloud distance threshold segmentation method performs best in the single-wood segmentation results of larch (average F-0.94), in which the multi-scale segmentation accuracy of young larch is comparable to the point cloud distance threshold method. The shape of the larch canopy is relative lys, the gap between adjacent trees is clear, and the height of the single wood tree is more average, the SFM method feature point matching accuracy is higher, so setting the appropriate point cloud distance threshold can achieve a better segmentation effect.

The result of evaluation index shows that the point cloud distance threshold segmentation method is the best in the single tree segmentation of Larch (average f = 0.94), and the multi-scale segmentation accuracy of young larch is equivalent to the point cloud distance threshold method. The crown shape of larch is relatively regular, the gap between adjacent trees is relatively clear, and the height of single tree is relatively average. The SFM method has a high accuracy of feature point matching, so setting a suitable point cloud distance threshold can achieve a better segmentation effect.