Red Edge Position Analysis for Detecting Vegetation Types and Age Using Hyperspectral Remote Sensing
要旨Two red edge position (REP) techniques, Linear and Lagrangian, were applied on hyperspectral data acquired from the HyMap sensor for a forested area in Thetford Forest, UK. Red edge positions of different vegetation covers were extracted with the two approaches from the hyperspectral data. Based on the estimated REPs, the Linear and Lagrangian interpolation methods were compared with ground reference image for analysis of different vegetation types and age. Experimental results of both interpolation techniques indicate that the wavelength and reflectance of REP for younger plants (higher chlorophyll content) shift towards longer wavelength and higher reflectance in comparison with older plants (lower chlorophyll content). The Lagrangian approach is more accurate than the linear method for estimating REP of grassland because it minimizes the soil background reflectance effects.