论文部分内容阅读
以LIDAR测量树高为研究内容,利用全站仪对研究区选定的37株样木的树高和材积进行精准测量,获取训练样本实测数据。通过BP神经网络分析,建立了LIDAR测量树高与实测树高、材积的拟合函数方程,实现了用LIDAR树高数据计算单木材积的过程。从拟合的结果来看,树高相关系数R为0.993 2,平均残差平方和为0.086 312。材积相关系数R为0.974 7,平均残差平方和为0.000 310 5,达到了林业调查精度的要求,在林业数表的编制方面是一个比较好的范例。
With LIDAR tree height as the research content, using the total station, the tree height and volume of the 37 sample trees selected in the study area were measured accurately to obtain the measured data of the training samples. Through BP neural network analysis, the fitting function equation of LIDAR measured tree height and measured tree height and volume is established, and the process of calculating single timber volume by LIDAR tree height data is realized. From the fitting results, the tree height correlation coefficient R was 0.993 2, and the average residual square sum was 0.086 312. The volume correlation coefficient R is 0.974 7, the average residual square sum is 0.000 310 5, which meets the requirements of forestry survey accuracy and is a good example in the compilation of forestry number tables.