基于低密度机载LiDAR和CCD数据的林分平均高提取

Forest Mean Height Extraction Based on the Low-density Airborne LiDAR and CCD Data

本文基于低密度的机载激光雷达(L iDAR)数据生成林区树冠高度模型(CHM),结合高分辨率CCD数码相机影像勾绘林分多边形,由改进的树冠识别算法提取林分平均树高。结果表明:全部有效数据林分总体精度达74.86%,刺槐精度达75.62%,油松精度达74.74%,结果受点云密度影响,使得阔叶树种的精度稍高于针叶树种,因此,低密度激光雷达数据结合高分辨率CCD可以快速、准确地提取林分平均高。

The study generated forest canopy height model （CHM） using the low-density arborne Light Detection and Ranging （LiDAR） , it sketched the stand polygons combined with the high-resolution Charge Coupled Device （CCD） digital camera image, and extracted the stand mean height by the improved recognition algorithm. The results showed that the overall accuracy of the total valid data was 74.86%, The accuracy was 75.62% for Robinia pseudoacacia, and 74.74% for Pinus tabulaeformis. The precision of broad-leaved was slightly higher than that of the conifer species owing to the influence of the point cloud density. Therefore, low-density LiDAR with high resolution CCD data can quickly and accurately extract the stand mean height.