The Recognition of Altered Rock Based on Spectral Modeling and Matching Using Hyperspectral Data

In this paper, we focused on the alunite alteration recognition which was characteristic of hydrothermal alteration zones in the study area of Cuprite mining area in southwestern of Nevada. Two solutions were practiced to improve the alteration recognition precision and reliability. One is the modeling of object reflectance spectra with the combination of linear and nonlinear model, which was taken as reference spectra instead of the mineral spectra from spectral library, another is the fusion of the information extraction methods including spectral characteristic analysis and spectral matching by the weights decision function. Several rule images were obtained in which each pixel representing the similarity between the corresponding pixel in the hyperspectral image to a reference spectra. An alteration map with improved precision was created after weights decision processing.