喀斯特山区草地地上生物量遥感估测

Estimation of Karst Mountain Grassland Aboveground Biomass Based on Linear Spectral Mixture Model

由于喀斯特山区地表破碎,环境异质性强,影像上混合像元现象严重,增加了草地地上生物量遥感估测的难度。以Landsat ETM+为数据源,将光谱混合分析技术应用于喀斯特山区草地地上生物量的遥感估测中,分析了基于线性混合分解模型分解的草地分量与实测草地地上生物量的相关关系,进而构建了喀斯特山区草地地上生物量估测模型为y赞=2 451.158x-280.461(r2=0.801 2),验证结果总体精度可达85%以上。基于光谱混合分析建立的草地地上生物量估测模型提高了喀斯特山区草地地上生物量遥感估测的精度,为喀斯特山区草地生长状况调查与监测提供了技术支持。

The grassland is an important part of the Karst mountain ecosystems in Southern China and the basis of protecting the karst mountainous ecological environment and developing of grassland animal husbandry. In typical karst mountain area,the complex and broken ground surface leads to serious mixed pixel phenomenon. Linear spectral mixture model was used to estimate grassland aboveground biomass based on Landsat ETM＋image.The results showed that the grassland fraction extracted with the linear mixed model had obvious lineal correlation with the field data measured. Grassland aboveground biomass estimation model can be expressed by the equation y^=2 451.158x-280.461（r^2=0.801 2）. The accuracy validation showed that the overall accuracy was above 85%. The estimation model of grassland aboveground biomass which based on spectral mixture model can improve the estimation accuracy in karst mountainous area. This method can provide technical support for surveying and monitoring the growth of karst mountainous grass.