林下植被遥感反演研究进展

The Progress of Forest Understory Retrieval from Remote Sensing

林下植被在森林生态系统碳、水和营养元素的累积和循环方面有着重要作用与科学意义。多角度、高光谱和激光雷达遥感系统凭借对森林分层结构的敏感性,成为量化林下植被的重要手段。本文综述了森林林下植被的遥感反演研究进展:首先讨论了林下植被的定义,其次对当前林下植被的遥感反演现状作了深入分析,总结了多角度、高光谱和激光雷达遥感观测的森林背景的反射率、林下植被的叶面积指数、高度和覆盖度的遥感反演原理和方法。基于不同卫星观测角度下森林冠层和森林背景对总反射的贡献差异,林下反射率可通过多个角度的观测数据进行反演。此外,借助激光雷达穿透冠层直接观测林下植被的优势,总结了激光点云数据和回波波形信息反演林下植被的覆盖度和高度的方法,以及今后使用遥感技术反演的难点和获取林下植被信息的主要发展方向。

Forest understory contributes significantly to cycles and accumulation of carbon, water and nutritional elements in the ecosystem. Remote sensing with multi-angular, hyperspectral or Lidar systems is very sensitive to the stratification of forest, which allows us to quantify understory. In this paper, the progress of forest understory retrieval from remote sensing is reviewed. The definition of understory is firstly discussed. Then, a systematical analysis is presented. The review focuses on forest background reflectivity, understory leaf area index, and understory height and coverage that are retrieved respectively from multi-angular, hyperspectral and Lidar data.Based on the different contributions of forest canopy and background with respect to different observation angles, multi- angle observations make it possible to retrieve forest background reflectivity. The retrieved forest background reflectivity could further be used to estimate understory leaf area index through various retrieval algorithms. In addition, as Lidar systems are capable of directly observing understory through forest canopy, understory height and coverage could be retrieved from Lidar cloud points and waveform information. Finally, the shortcoming of current research and possible improvements in future research are discussed. There are mainly three facets where future understory retrieval could emphasize, including the using of spectral, phonological and angular differences between understory and overstory, the simulation of multi-angular observations by bidirectional reflectance function, and the fusion of multi-source data from radar, Lidar and hyperspectral remote sensing systems.