Lithological Identificaiton And Extraction Of Granitoids From ASTER Multispectral Data

Remote sensing petrology is a significant aspect of remote sensing technologyapplication on geological science. Lithological information can be detected and identifiedby using remote sensing data based on spectral characteristics of rock and mineral.Granitoids, as the momentous rock type of igneous rock, its lithological identificationand extraction is one of hot topics in remote sensing petrology because of the tectonicimportance and its relationship with polymetallic ore deposit origin. The western Junggaris located on the northwestern part of Xinjiang, is an important metallogenic area inChina, and a number of granitic plutons are well exposed in this region because of thesparse vegetation with the arid climate. Therefore, it is an ideal natural laboratory tostudy the identification and extraction of granitoids from remote sensing data. Theobjective of this thesis is to establish a set of methods for extraction and identification onlithological information of granitoids in the western Junggar using ASTER multispectraldata based on the spectral measurement and analysis of spectral characteristics of rocksamples. Furthermore, classification of granitoids and mineral assemblage in the studyarea will be discussed combined with the result of data processing and interpretation.This thesis make an attempt to provide scientific grounds for geological survey andmetallogenic exploration in this area，and a set of methods for the identification andextraction of granitoids in the arid and semi-arid region.Major results and conclusions are as follows:（1）Granitoids rock samples collected from the western Junggar generally has notobvious spectral features in the VNIR-SWIR region, and there is the small variability indifferent granitoids. Some absorption bands from iron ion converge on VNIR region.Compared with plagiogranite, iron spectral absorptivity of alkali-feldspar granite isstronger. The variation of hydroxy absorption bands in SWIR region are as follows: TheAl-OH absorption band of monzonitic granite is centered at2.2μm （ASTER band6）because of rich muscovite and sericitization alteration；The Mg-OH absorption band canbe observed at2.35μm （ASTER band8）caused by biotite rich in alkali-feldspar granite;The hydroxy absorption bands are located at2.25μm and2.38μm respectively.（2）Compared with the731optimal bands combination based on the optimal indexfactor (OIF) calculation, the principal component analysis (PCA) processing of nine ASTER VNIR-SWIR bands is more effective. The granite rocks present brighter colorand distinguished feature in the false color composite images of PC245. For example,four types of grantiods in the Akebasito pluton can be identified successfully throughinterpretation and analysis. Considering the enhancement of band ratio, the selected ratioof2/1,4/8and relative absorption-band depth (RBD)6are employed to indentifylithological information of alkali-feldspar granite, monzonitic granite and plagiogranite,respectively.（3）Spectral signatures of quartz and feldspar in thermal infrared region areobviously stronger than that in VNIR-SWIR region. The spectral curve of rock samplespresent synthetical features of quartz and feldspar. The emissivity of alkali feldspar andplagioclase at ASTER band10and12is remarkably higher than that at band11.However, the quartz shows the opposite feature. At plagioclase groups, the spectrtal slopebetween band12and band13is declining from endmember component of albite toanorthite. In contrast，the emissivity of alkali feldspar at band13is generally higherthan that at band12.（4）According to the analysis of spectral characteristics, quartz index(QI), alkalifeldspar index (AKI) and plagioclase index (PLI) from ASTER TIR data can be able tomap the distribution characteristics of quartz, alkali feldspar and plagioclase in granitiods.Meanwhile, threshold segmentation of these three indexs and mafic index (MI3) revealthat lithology of three plutons in the study area are alkali feldspar and plagioclase, whichare mostly located in the central and western part of Akebasito pluton and thenortheastern part of Hongshan pluton. Alkali-feldspar granite, plagiogranite, monzoniticgranite and ultramafic rock can be distinguished semiquantitatively. Mixture TunedMatched Filtering (MTMF) technique is applied to detect composition of quartz andfeldspar in different granitoids. The dominant mineral is Albite, Perthite and Orthoclasesecondly. These results indicate that alkali-feldspar graniteis is typical rock in study area.（5）False color composite image of AKI、RBD6and PLI fromASTER SWIR-TIRdata fusion can effectively be applied to distinguish the different granitoids in study area.The alkali-feldspar granite, granite, granodiorite and monzogranite can be successfullyidentified. Furthermore, ASTER imagery interpretation of lithological and tectoniccharacteristics also suggest that the partial areas in Ⅱ pluton and the interchange betweeneastern I and III plutons of Karamay are potential region for the metallogenicexploration.