研究目的
To detect and analyze the health of individual trees in urban environments by integrating LiDAR and hyperspectral data for geometric and physiological parameter extraction.
研究成果
The method effectively combines LiDAR and hyperspectral data to extract individual trees and compute geometric parameters with high accuracy (errors below 2.08% for key parameters). It identifies trees needing attention based on height and inclination. Hyperspectral data provides potential for physiological health analysis, but further study is required for comprehensive health evaluation.
研究不足
Lack of real measurement data for accuracy validation of geometric parameters; reliance on simulated trees for error calculation. The in-depth analysis of physiological parameters from hyperspectral data is not fully explored and is deferred to future studies.
1：Experimental Design and Method Selection:
The methodology involves a multi-sensor platform combining terrestrial LiDAR and hyperspectral data. The process includes calibration, individual tree extraction (via ground removal, Euclidean distance clustering, and NDVI-based filtering), and health analysis (computing geometric parameters like height, DBH, crown diameter, inclination from LiDAR, and physiological parameters from hyperspectral data).
2：Sample Selection and Data Sources:
Data was acquired in September 2016 on Huandao Road, Xiamen, focusing on camphor trees. 175 trees were extracted from the point clouds.
3：List of Experimental Equipment and Materials:
RIEGL VMX450 mobile laser scanning system for LiDAR data, Headwall's Nano-Hyperspec for hyperspectral data in the VNIR range (400-1000nm).
4：Experimental Procedures and Operational Workflow:
Steps include sensor registration, data preprocessing (cutting and ground removal), clustering, filtering using NDVI, and parameter computation. Accuracy was validated using simulated trees from OnyxTree software.
5：Data Analysis Methods:
Statistical analysis for parameter errors (e.g., height, DBH, crown diameter errors calculated as percentages), and use of vegetation indices like NDVI for health assessment.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
