The Study On Remote Sensing Lithologic Information Mapping Method In Carbonate Terrane

There are about 3,400,000 square kilometers of carbonate terrane in China, which account for 35.41% of the whole land territory. The carbonate terrane distributes mainly in eight provinces in southwest China, which include Guizhou, Yunnan, Guangxi, Chongqing, Sichuan, Hunan, Hubei and Guangdong. Though several times of geologic survey based on lithologic mapping have been done since 1950s, yet the progress of investigation is getting slowly because of the field survey method. Problems of resource exploration and ecological protection are put out with the rapid and sustaining economic development, so, an urgent need for Remote Sensing technology is put forward to accelerate geologic lithologic survey and improve the survey quality, which will speed up the pace of exploiture of resources and reconstruction of ecosystem. At present, less study have been done on Remote Sensing lithologic identification in southwest China, and little mature experience can be followed. So, it is very difficult to identify lithologic units from Remote Sensing image because of the thick soil and flourishing vegetation.In this paper, theoretical base of identify lithologic units in carbonate terrane be offered by elaborating the mechanism of rock spectroscopy, studying the spectral features of carbonate rock, and analyzing the differences of spectral feature between dolomite rock and calcite rock. Based on the field survey, a piece of lithologic section be measured, field lithologic points be set up, spectral curve(1300~2500nm) of rock samples be measured, and the lithologic spectral library of the study area be founded. At the same time, Remote Sensing lithologic units of the study area be divided according geology data and lithologic spectral features.Multispectral Remote Sensing technology mainly identify lithologic units by spectral feature of rocks. But in the diagnostic spectral range of carbonate rocks (2000~2350nm), multispectral data often only has one band, for example ETM only has band7. It brings inconvenience to lithologic identification because of the long spectral range of one band and the image pixels which often mixed by the information of vegetation, soil and rock. In the paper, the image features of lithologic stratum be studied, then image signs be set up, and lithologic stratum can be identified by manpower. As one conclusion, it gets good result by automatic classifying both Remote Sensing image and texture image, then amending the classified image with computer based on three-dimensional Remote Sensing image model and lithologic stratum distributing rules. Except spectral feature, rocks have difference in aspect of terrain and relief, such as height and slope of hill, performance of soil and character of water system, etc. The terrain and relief of dirrerent lithologic units be studied in the paper, and image signs be set up with IKONOS image, then lithologic units be identified by method of manpower. In conclusion, with high spatial image, carbonate rocks and un-carbonate rocks can be identified very well, and calcite and dolomite rock can be identified.Hyperspectral image is the best data when carring on lithlologic mapping by Remote Sensing technology. With the MASTER image of Cuprite Remote Sensing geologic study area, many kinds of mineral's information be identified. The author proves it is possible to map carbonate lithology in bared area with hyperspectral data by comparing the minerals'information with the study result which other scholar did in the same area. And carbonate lithologic units be identified by method of spectral angle mapping with Hyperion data of ShiTang–ChaoNan study area.