融合LiDAR点云与高分影像的单木检测方法研究

Research on the Individual Trees Detection Method Based on LiDAR Combined with High-resolution Image

对一种结合高分影像数据和机载LiDAR点云的单木检测方法进行研究,首先采用面向对象分类方法对高分影像上的单木区域进行分割,得到单木区域分割图;再以单木区域为约束,从机载LiDAR激光点云中分离出单木点云,构建局部冠层高度模型(CHM);最后对单木CHM采用分水岭算法探测冠层局部极值,实现对单木中心定位检测。实验结果表明该方法能够充分利用高分影像和激光雷达点云优势,提高了单木定位准确性,具有广泛的应用前景。

In this paper,we proposed a method combining high-resolution image and airborne LiDAR point cloud.Firstly,the individual trees’area on the highresolution image was segmented by object-oriented classification method to get the map of individual trees.Secondly,the individual trees’point cloud was separated from the airborne LiDAR point cloud with the individual trees’area as the constraint,and the canopy height model(CHM)was constructed.Finally,the watershed algorithm was used to detect the local extreme value of the canopy for the individual trees’CHM to realize the individual trees’center and location detection.The experimental results show that this method can make full use of the advantages of high-resolution image and LiDAR point cloud,improve the accuracy of individual trees’positioning.The accuracy of individual trees’detection is better than 95%,and the accuracy of individual trees’positioning is better than 0.5 m,which has a wide application prospect.