The Research Of Rock Mineral Information Extraction Based On Hyperspectral Remote Sensin In Tibetan Plateau

Hyperspectral Remote Sensing technique is a very important part of remote sensingscience for its integrating of images and spectrums, and it plays a major role inquickly acquiring huge quantity geographic information at the present stage. Regionalgeological survey and mineral resources investigation indicate the level of geologicalresearch work in a country or region. Because of the rapid development of geologicalsurvey and depth research in geological surveys, Hyperspectral Remote Sensingtechniques and its applications earn widespread respect in this field. Tibetan Plateau isa high altitude area and it has rich mineral resources. It plays an important role forChina’s energy strategy. This paper carried out the research on remote sensing datacollection, data processing, information extraction and recognition method andapplication of rock and mineral. The research results of this paper can supportgeological remote sensing techniques. The main achievements and contributions areas follows:The paper designs a set of minal and rock information extraction methods toTibetan Plateau with hyperspectral remote sensing. It discusses the features of rockand mineral spectrum and their response mechanism, and the typical absorption ofmineral ions and the characteristics. Using the Hyperion, ASTER, ETM, IRS-p6andCBERS-02b remote sensing data, successful accomplish identification of a number ofregional typical rocks and minerals. Through sample testing and analysis in laboratory,effectively verify the validity of information extraction and recognition. Finally, basedon practical application, some suggestions are given for hyperspectral remote sensingtechnology in modern geological mapping. The paper researches the typical rock and mineral spectral characteristic anddiscoveries of the spectrum pressure effect to Tibetan Plateau. Spectral measurementis carried out as a pilot area in the East Kunlun orogenic belt, including limestone,sandstone, conglomerate, marble, dolomite and other categories. Under naturalconditions, combined with the weathering characteristics of regional land cover,which is covered by loess, the measurement of spectral response is carried out. Basedon this analysis, the paper gives rock spectrum diagnosis in typical location. Inaddition, the spectrum pressure effect of plateau region is found. As Tibet’s highaltitude, low air pressure, water vapor content little, the atmospheric absorption bandcenter has been offset, width narrow. The discovery is very important to carry outremote sensing basic research and application research.This paper designs a spectral matching method for information extraction to rocksand minerals.The process is designed based on the minimum distance method, thecorrelation coefficient method and the classification window, taking into account theeffect that the instability and uncertainties of environmental factors bring to the rockore, and this method is a targeted integrated spectral matching method. The idea isbased on assuming that environmental factors impact on the rock is random(temperature, humidity, water vapor, air etc), and supposing the characteristics is anormal distribution. And in the whole wavelength range, the curve of spectrum isdivided to the spectral characteristics of range based on the effective contribution,total whole window coefficient of correlation is carried out, and the ultimatecorrelation coefficient is a reference for rock and mineral identification.