Extraction And Analysis Of Information Of Alteration With Multi-source Remote Sensing Data In High Vegetation Coverage Area

Multi-source remote sensing data usually are applied to extract mineralization-related alteration information to achieve the delineation or prediction of favorable mineralization targets in the exploration of mineral resources.The surface vegetation interference to mineralization information extraction in the vegetation cover area will have a great impact on the accurate delineation or prediction of the favorable target area.The research area located on northern area of Shibing County in Guizhou province.The mineral resources are relatively abundant in the area,however,extracting alteration information is harder in the higher vegetation coverage area with remote sensing data.Focused on the above problems,both methods as vegetation coverage + mask and mixed pixel decomposition were selected to reduce the vegetation influence in the research.ASTER and Hyperion data were preferred to extract the mineralization alteration information,respectively.Further,The results based on the two methods were compared and analyzed to select the better method for extracting the alteration information in such area,and then scientific reference for the extraction of remote sensing alteration information could be provided in remote sensing fields.The main work of this dissertation is summarized as follows:(1)Preprocessing ASTER and Hyperion data based on digital image processing method.The FLAASH module of ENVI software was used to carry out the atmospheric correction of multi-spectral and hyperspectral data.The corrected image data can not only better reflect the reflectivity information of the ground objects,but also present the higher clarity of images.(2)Using ASTER multispectral data to extract iron and carbonate alteration information from the study area.According to the characteristics of vegetation coverage in the study area,the vegetation coverage + mask and mixed pixel decomposition were selected to eliminate the vegetation effect.Further,the alteration information was extracted with the ratio method and the principal component analysis method,respectively.With the superposition analysis and field verification of previous geological data,"mixed pixel decomposition + principal component analysis" was proved not only to inhibit vegetation sufficiently and highlight mineralization information,but also to have the higher accuracy of extraction result and be highly related to stratum and fault in spatial distribution.(3)The mineralization of carbonate in high vegetation coverage area was studied with Hyperion hyperspectral data.Focus on the characteristics of hyperspectral data,the Kernel Principal Component Analysis method was applied to reduce its dimension,and then the better results were obtained.Moreover,PPI and N-dimensional visualization tools were taken to extract the pure end member,and then the matched spectral analysis were proposed between the pure end member extracted and the measured data.Finally,the spectral angle technique was used to carry out the mineral mapping.The results show that the extraction results of dolomite and calcite information from hyperspectral data match the extraction results of carbonate alteration information from multi-spectral data,while,the extraction of calcite and dolomite from hyperspectral data is clearer than that from multi-spactral data.