基于几何光学模型反演林冠郁闭度的普适性方法研究

Monitoring forest structure by quantitatively deriving canopy structural variables, such as crown closure, would enable further understanding and assessments of changes in forest ecosystem services and provide insight into processes that can be of influence. This research focuses on the inversion of physical based Li-Strahler geometric-optical model for estimating the forest crown closure at regional scale. The pixel-based input parameters will be evaluated individually, which influences the accuracy of the model inversion. In addition, a regional scaling-based method will be extracted that can upscale the information from higher spatial resolution data to lower spatial resolution data, and better derive the inputs of the lower spatial resolution image for the model inversion. The inverted Li-Strahler model combined with the scaling method makes it possible to estimate forest crown closure in large areas. By using multi-temporal data, a change detection in forest and the evaluation on forest ecological benefit can also be carried out.

针对定量遥感技术在森林监测上的需求，构建基于几何光学理论的林冠郁闭度逐像元反演模型，提出基于多源遥感数据的区域性空间尺度转换方法，根据林冠类型、分布特征、数据尺度效应对反演模型的输入参数进行优化与敏感性分析，解决利用物理学模型在大区域开展长时间序列森林动态监测的普适性问题，实现科学、合理的区域森林生态效益评价。

本项目针对定量遥感技术在森林监测上的需求，利用高光谱、高分辨率及激光雷达数据，结合地面调查，开展森林冠层参量的估算方法研究。设计完成了研究区机载高光谱和激光雷达LiDAR航飞及数据处理；在研究区获取森林冠层郁闭度及多种冠层参量地面实测数据；针对几何光学模型反演郁闭度的输入参数进行敏感性分析，优化了林冠郁闭度逐像元反演方法，构建基于几何光学理论的林冠郁闭度逐像元反演模型；利用激光雷达数据与多时相光学遥感数据在典型研究区以及大范围研究区两个尺度开展林冠参量的估算与验证，提出基于多源遥感数据的区域性空间尺度转换方法，完成了三峡库区的森林郁闭度动态监测与分析。在项目执行期间，以项目第一标注发表SCI文章2篇，EI文章1篇；参加国际学术会议6次，国内学术会议2次；培养博士研究生2人，硕士研究生1人；获得科技支撑项目子课题1项。