Study on the Intelligent Extraction Based on Hyper-Spectral Characteristics of Typical Vegetation

By using hyper-spectral remote sensing data of desert vegetation, the original spectral data was simply pretreated firstly, then first order differential transform and smoothing was the hyper-spectral data. The spectral characteristics of different grassland types were extracted. The results showed that: desert vegetation has some unique spectral features of common green vegetation. However, affected by the underlying surface of spared leaves, low coverage, the spectrum of desert vegetation does not have obvious green peak, and the red edge characteristics decreased with the decline of vegetation coverage.