A multiple-stage algorithm to separate hydrothermal alteration zones by ASTER satellite data: A case study from Kerman province

Image processing of remote sensing for separation hydrothermal alteration, in the case of missing initial spectra of pixels, can be a challenge for the researcher. The previous researches has shown that accurate separation of hydrothermal alteration zones using Conventional methods of image processing based on Spectral Properties of pixels, is not possible. Therefore, this research is trying to present a multi-stage algorithm that identify and discriminate the hydrothermal alteration zones in western part of Kerman province with high accuracy. To achieve this goal, the principal component analysis method, Fractal Concentration-Area Model and Full Index Kriging (FIK) geostatistical model are used in combination. The results show the high accuracy of the FIK model in identifying and separating each of the phyllic, argillic and propylitic alterations in the study area. Also, to evaluate the classification error, the Confusion matrix was investigated. The results of the Confusion matrix showed that the FIK model performs well in terms of image classification. Also, given the high number of training pixels in the phyllic zone, the FIK model has been able to identify this type of alteration very well.