基于机载LiDAR落叶松树冠几何形状三维重建关键技术研究
Key Technologies of Three-dimensional Geometric Reconstruction of Larch Crown Shape Based on LiDAR Data

森林冠层的三维重建研究能够更加直观反映森林空间结构，提高森林参数的测量精度。目前小光斑激光雷达已经广泛应用于林业研究中。为建立落叶松树冠三维形状模型，以长春净月潭实验区落叶松机载LiDAR（LiDAR, Light Detection And Ranging）数据为基础，采用K-means算法提取建模参数。该算法以单木树冠顶点作为初始聚类中心，经过4次迭代估测出单木树高和单木树冠直径，通过与试验区的单木实测数据对比，进行相关性分析，得到估测树高和估测树冠与实测数据相关系数分别为0.892 4和0.769 0，经过验证，估测树高和估测树冠的精度为94.06%和82.21%。利用激光雷达提取出的单木坐标、树高、树冠和冠基高采用旋转抛物线方法重建森林尺度三维模型呈现森林结构。
The research on forest three-dimensional reconstruction can response forest structure intuitively and improve estimation precision of forest parameters. At present, the small footprint LiDAR has been widely used in forest study. To establish the three-dimensional model of larch based on its canopy shape, the field measured data of larch at Jingyuetan area in Changchun and airborne LiDAR data were selected，and the K-means cluster algorithm was used to extract the modeling parameters, in which the individual tree top was taken as the initial clustering center, and individual tree height and crown diameter were estimated after four iteration procedures. Regression analysis were made between the field measured individual tree height and crown diameter and LiDAR measured individual tree height and crown diameter, resulting in the corresponding correlation coefficients of 0.892 4 and 0.769 0, and the precisions of 94.06% and ？？82.21%？？, respectively. Finally, the 3D model to represent forest structure was established by using the parameters such as the LiDAR measured individual tree top coordinate, tree height, crown diameter, crown base height and a rotating parabolic method