基于线性光谱分解方法的森林选择性砍伐与再生监测研究- 以巴西亚马逊森林为例

以巴西亚马逊热带森林为例,在Landsat TM、ETM+影像提取光合植被(PV)、非光合植被(NPV)、土壤(soil)端元光谱的基础上,利用线性光谱分解模型得到端元地物丰度,进而设计合理的检测规则实现了选择性森林砍伐与森林再生所产生的森林变化监测。结果表明,基于线性光谱分解的方法能够以较高精度准确检测出选择性砍伐与森林再生,是森林变化宏观监测的有效手段之一。

Monitoring selective logging and forest regeneration based on linear spectral mixture method: A case study of the Brazilian Amazon

Taking Amazon rain forest in Brazil as an example, this study selects the endmember spectrum of photosynthetic vegetation (PV), non-photosynthetic vegetation (NPV), the soil from TM/ETM+ images, and uses linear spectral mixture model to retrieve the fractional cover of PV, NPV and soil. In addition, the paper discusses the detection rules used to detect the selective logging and forest regeneration. Results show that there is a higher precision for detecting selective logging and forest regeneration based on the linear spectral unmixing method, which can be effectively used to monitor the macro-scale forest change.