Determining the Optimum Position of Boreholes, Using Hyperion Image and Neural Networks Method

Among detailed mineral exploration studies, the alteration mapping has been proved to be the fundamental objective for identifying the deposits formation. This paper aims determining alteration zones in Hyperspectral images with the assist of SAM & Band Ratio methods at first. The second interest is integrating the output of determined alteration zones by other mineralization factors using different Neural Networks namely Multilayer pecreptrons, Radial Basis Function & Generalized Neural Network and using cross correlation method. This integration is performed for determining the position of boreholes of porphyry copper exploration in Nowchoun region. In the case of Hyperspectral classification, the best result have been achieved by the band ratio method, i.e. about 94.2 percent. Eventually, the degree of correlation between maps that produced by neural networks and operated exploration boreholes have been estimated. Comparison between the high potential points indicated by our maps with those previous drilled boreholes reveals that MLP network has the highest correlation. This correlation is about 54% in Nowchoun region.