Remote Sensing Research On Extraction Of Rocks And Altered Minerals Information From Huangshan Area In Xinjiang Using ASTER Data

Research on spectral characteristics of rocks and minerals is the basis of remotesensing petrology. There are two atmospheric window which is suitable for the studyof petrological and mineralogical characteristics.:0.4~2.5(the visiblenear-Infrared and short wave-length-infrared band) and8~14(the thermal infraredband). They embody the reflection and emission characteristics. Therefore, combininguse of the two parts to establish extraction process will improve the ability of mineralsidentification.In the paper, rocks and minerals information extraction is topic contents, based onthe ASTER multispectral data. Two aspects research are carried out as follows:mineral alteration information extraction with the visible-short wave infrared remotesensing and the rock ore oxide content with the thermal infrared remote sensing.verification.on the basis of Regional geological background and field work, the papermainly made the following results.(1) PCA(Principal Component Analysis),"RBD(related band depth)+PCA"combination analysis method and "MNF+PPI+N-dimensional visualization+SAM"hyperspectral mapping process are applied in mineral alteration information extraction.Based on the verification work, conclusion can be obtained. In kinds, the number ofMethod3is six, included chlorite+epidote, muscovite, dolomite+calcite, tremolite+amphibole, chlorite+antigorite, chlorite. From the point of range, mineral alterationinformation overlay more in Method1, meanwhile Method2distinguish Mixed pixelbetter. However, hyperspectral mapping process has advantage of precision evaluation.Every method has gains or losses. A cross-validation study can be carried out usingthe method above.(2) In the research of rock ore oxide content extraction, nonlinear fittingprocesses are analyzed with mineral emissivity band ratio and chemical component in ASU thermal infrared spectral library. Inversion model draw a conclusion of thecontent of SiO2, Na2O, K2O, MgO. Research shows SiO2content is retrievedquantitatively to identify the acidity of the rocks; MgO content indicate the position ofdark-colored femic minerals; RittMann Index(calculated by SiO2,Na2O,K2O content)have a great indication function of acidity and alkaline.