Full unmixing hydrothermal alteration minerals mapping by integration of pattern recognition network and directed matched filtering algorithm

Partial unmixing categorizes each pixel based on a particular mineral while a pixel may contain several minerals. Hence, it seems that the full unmixing procedure is essential to make real results. In this study, at first, the hydrothermal alteration minerals have been mapped by directed matched filtering (DMF) algorithm as a partial unmixing method on the most important porphyry copper deposit in the Kerman province in Iran, then the full unmixing procedure was performed based on a pattern recognition network. In fact, the pattern recognition network uses results of the DMF algorithm to measure the amount of each alteration mineral in each pixel. According to the achieved results, the pure pixels of alteration minerals show high mixing with each other so that the average purity of kaolinite, muscovite, chlorite and alunite pixels are 80%, 70%, 95%, and 92%, respectively. Moreover, the spectral results of 30 samples and 212 chemical analysis of field samples by other research in this area validated those obtained by the full unmixing procedure.