Proc. SPIE. 9142, Selected Papers from Conferences of the Photoelectronic Technology Committee of the Chinese Society of Astronautics: Optical Imaging, Remote Sensing, and Laser-Matter Interaction 2013
The first-order derivative transformation was applied on a PROBA CHRIS hyperspectral remote sensing image of the Yellow River Estuary coastal wetland. Five classic supervised classification methods were employed on the images before and after the derivative transformation, and then those classification results were compared through manual interpretation and quantified analysis. The aim of this research is to evaluate the effects on the classification ability of supervised classification methods made by the derivative transformation. Experimental results show that, the derivative transformation is capable of improving the classification ability of certain supervised classification algorisms in coastal wetlands classification using hyperspectral images. Especially, for the Maximum Likelihood and Support Vector Machine methods, with the best classification accuracy, derivative transformation could effectively help distinguish vegetation and clear water wetlands.