Assessment of individual tree detection and canopy cover estimation using unmanned aerial vehicle based Light Detection and Ranging (UAV-LiDAR) data in planted forests

本文考察了三种郁闭度估算方法的效果。1) 基于单木分割提取法；2) 基于CHM提取法；3) 基于特征变量模型法。本文特征变量模型预测法是通过逐步回归法构建的线性方程预测森林郁闭度，筛选出相关性高的变量，剔除相关性低的变量，最终得到最优模型。

This study evaluates the performance of three forest canopy closure estimation methods: 1) individual tree segmentation-based extraction method; 2) CHM-based extraction method; 3) feature variable model-based method. The feature variable model prediction method in this study constructs a linear equation for forest canopy closure prediction via stepwise regression, selects variables with high correlation, eliminates those with low correlation, and finally obtains the optimal model.