عنوان مقاله [English]
The classification of remote sensing images is one of the important parts of the image analysis. This study aims to extend the results of Hyperion data processing to ASTER images of Lahrud area for achieving a broader mineralogical distribution map. This method uses the high accuracy Hyperion data as training data set and high coverage ASTER scene for classification purposes. Because of the differences in spectral resolution between ASTER and Hyperion data, the ability of ASTER data to identify the training classes was evaluated and the separability score for the selected areas were calculated. In this study, ten minerals detected by the Mixture Tuned Matched Filtering (MTMF) method on Hyperion image were therefore used as training classes for separability computation. The classes with high similarity (low separability) were combined (six classes were distinguishable). The classification was then performed using three parametric methods of maximum likelihood, minimum distance, mahalonobis distance. The accuracy of the results of each classifier was evaluated by constructing the related confusion matrix. The matrix was computed using the consisted pixels in each training class before and after classification. The maximum likelihood revealed best performance in comparison to the other methods. The abundance maps of the minerals resulted by the mentioned method revealed a high coincidence with geological reports of the area. The widespread sulphate minerals (Gypsum and Polyhalite), the minor occurrences of Malachite around zeolite minerals as a narrow strip, and invasive presence of zeolite minerals like Analcime and Mesolite all are compatible with the geological literatures of the region.