Abstract:

In this study, the Seriphidium transiliense desert grassland as the research object and based on HJ-1A HSI hyperspectral image and field canopy spectrum, the identification methods of degradation class were studied using SAM and SID methods. The results showed that the identification accuracy of the canopy spectrum combined with HSI was poor, but the overall identification accuracy of SAM and SID methods were above 76% based on the HJ-1A HSI. It proved that the identification approach based HSI hyperspectral image was better than the canopy spectrum combined with HSI hyperspectral image.

Key words: HJ-1A HSI, Hyperspectral image, Desert grassland, Degradation class, Identification

摘要：

本研究以伊犁绢蒿（Seriphidium transiliense）荒漠草地为研究对象，基于环境卫星HSI高光谱影像和群落冠层光谱，采用光谱角填图法和光谱信息散度法对草地退化等级进行识别。结果表明：以实地采集的冠层反射光谱为指导的HSI高光谱影像识别精度较差；基于HSI高光谱影像的退化等级识别结果较好，总体分类精度在76%以上，适合对伊犁绢蒿荒漠草地退化等级识别。

关键词: HJ-1A HSI, 高光谱, 荒漠草地, 退化等级, 识别