Hyperspectral processing on Gachin dome Satellite data, using SFF and MNF to separate geological unites

Identification of clay minerals and associated alteration is considered as an important step in prospecting for the hydrothermal deposits, and using hyperspectral sensors ability is instrumental in this regard. Hyperspectral images of Hyperion sensor are a rich source of information with 242 narrow contiguous spectral bands. Of course there are several mechanisms which contaminate the spectral information content of some bands. Identification and removal of the spectral defects is an important issue in hyperspectral remote sensing. MNF method or Minimum Noise Fraction is anoxia rotation in a direction that minimizes the correlation between data. MNF transformation is composed of two separate PCA rotations that data beyond the mean plus or minus multiplied standard deviation of the histogram of pixels as identified malformed data with a standard PC conversion is cleared.in order to investigate the effect of noise reduction in hyperspectral processing. In this study, at first MNF rotation has been performed on the hyperion data of Gachin region and then based on the provided chart, the data below 10% are rejected and by turning the axis again, reset rotation axis of MNF driven to the first state. A case study was performed on Gachin salt dome and in order to study field results, sampling in different parts of the range was performed and the samples were studied on a manual and microscopic scale, as well as XRD and electromagnetic spectrometer. Investigating the results of processing methods indicates a positive effect of noise reduction in processing methods. Thus, the method of spectral Feature fitting (SFF) method reduces the noise reduction of the factor coefficient from 70% in the MNF image to more than 86% in the MNF image.