| The airborne Light Detection and Ranging(LiDAR) technology integrates the laser ranging technology, computer technology and positioning technology, could provide a brand new means to acquire information with high spatial and temporal resolution, which is a unique advantage in forest resources survey. The purpose of this paper is obtaining parameters of individual trees from the LiDAR data and digital photography data with high spatial resolution. This paper is carried out as follows:(1) It processes the de-noising, filtering and classification to the point cloud data, and educes the CHM with subtracting the DEM and DSM obtaining from the point cloud data.(2) It uses the morphological operation to smooth the null value in the CHM, and obtains the DOM by the ortho-rectification of digital image referring to the DEM in the same area.(3) Taking the optimized CHM and DOM as objects, this paper obtains individual trees parameters through two segmentation methods, watershed segmentation and object-oriented multi-scale segmentation respectively. After analysis and comparison of the results, it is knowable that the object-oriented method is prone to over-segmentation, while the watershed algorithm was prone to under-segmentation in the forest of high canopy density.(4) By overlapping the crown polygons and CHM data and spatial domain analysis, it confirms the maximum value of CHM in the same crown polygon as the height of tree, and the concentric circle of polygon the as tree’s crown. Through the correlation analysis of parameters between extracted values and field measured values, it finds that the determination coefficients of crown diameter and tree height are 0.349 and 0.749 respectively.(5) This paper developed a "Point Cloud and Image Data Processing Program" based on the data of point cloud and digital image. This program implements the display and processing of point cloud data, and adds PCL into it to lay the foundation for the next step of program development. |
PDF Full Text Request