Leaf Area Density Inversion For A Single Tree And Broadleaf Forest Based On Terrestrial LiDAR Data

Leaf area density is an important parameter to characterize the vertical distribution of leaf area inside the canopy. The exact inversion of leaf area density distribution profiles is of great significance to the study of vegetation carbon and nitrogen cycles, primary productivity and biomass estimation. As a new technology of surveying and mapping, the terrestrial lidar has small spot and peediness, high efficiency, strong penetrability, high density, and other characteristics, which can provide valid data information for the inversion of leaf area density of vegetation canopy. How to inverse the leaf area density quickly and accurately based on terrestrial Li DAR data has become a hot topic in present. We selected a kind of broadleaf tree(magnolia tree) in the campus of University of Electronic Science and Technology of China(UESTC) for the study. With the point cloud data and true color images obtained by terrestrial lidar, we introduced a fast method of leaf area density inversion in this paper. The main work and conclusions are as follows:(1) Data acquisition and pre-processingWe researched on the method of experimental data collection outdoor using the Leica Scan Station C10 laser scanning system based on the basic theory of three-dimensional laser scanning and acquired the experimental data of single tree and broadleaf forest in the area. Combined with the performance characteristics of the scanner, we also researched the registration method of multiple point cloud data from different azimuthal positions and the denoising method.(2) Leaf area density inversion for a single treeWe researched on the method of separating the leaf point cloud and stem point cloud using Difference of Normals(Do N) operator and had an attempt. The result showed that for a simple single tree the method has certain applicability. We studied on the voxel-based canopy profiling(VCP) and the construction of three-dimensional voxel model. The contact frequency, the correction coefficient for leaf inclination and the vertical projection area were computed for every horizontal layer thickness and thus extracted the leaf area density distribution of the single magnolia tree. It was found that voxel size and the layer thickness were two important foctors affecting leaf area density inversion.The vertical distribution of leaf point cloud and leaf area density extracted based on the condition that the voxel size was 0.05 m × 0.05 m × 0.05 m and the thickness was 0.5 m was consistent. The LAD values tended to increase with the increase of the height at the middle and lower canopy and the maximum was 0.88 m2/m3 at the height of 3.5 m while the LAD values tended to decease with the increase of the height.(3) Leaf area density inversion for broadleaf forestWe researched on the method of extracting leaf point cloud by using the image classification and colouring point cloud based on the true color photo acquired by the Li DAR system and had an attempt. The result showed that this method was suitable for the extraction of broadleaf forest leaf point cloud. We inversed the leaf area density of broadleaf forest leaf point cloud. The result showed that the vertical distribution of leaf point cloud and leaf area density extracted based on the condition that the voxel size was 0.1 m × 0.1 m × 0.1 m and the thickness was 0.5 m was consistent. The LAD value tended to increase with the increase of the height at the middle and lower canopy and the maximum was 1m2/m3 at the height of 1 m, while the LAD value tended to decease with the increase of the height. Leaf area index calculated based on the leaf area density was 3.20 m2/m2 and it was 0.04 m2/m2 higher compared with the data derived from LAI-2200.