Prediction Of Metallogenic Prospective Area Based On The Fusion Of Multi Remote Sensing Data

Remote sensing technology has a unique advantage on mineralization and alteration abnormal information extraction, fault information extraction, and predication of prospective areas. Some progress has been achieved in geological prospecting in use of multispectral and hyperspectral data, such as Landsat TM, ASTER and Hyperion data. Unfortunately, the spectral resolution of multispectral data(e.g. Landsat TM, ASTER) is insufficient, even it has a high spatial resolution, making the extraction of alteration abnormal information be limited. The hyperspectral data of Hyperion’s spectral resolution is very high, while the amplitude and width is not suitable for a wide range of large area prospecting. According to their respective limitations, El Salvador of northern Chile(geographic range: 26°11’25″~26°29’51″S, 69°21’24″~69°40’57″W) is chosen as the study area and the spectral data have been reconstructed by fusion processing of ETM+ to Hyperion and ASTER to Hyperion in use of the spectral resolution enhancement method(SREM) proposed by Sun Xuejian, receiving wide coverage of high spectral resolution images. On the basis of this, the fracture structure and alteration anomaly information of the study area have been extracted and analyzed. Finally, based on optimi zation algorithm of evidence weight, a preliminary attempt is made to study the prediction of metallogenic prospective area in use of multi-source remote sensing data. The results of the study include the following three aspects:(1) Extraction results of alteration anomaly by Hyperion hyperspectral data after fusion to ETM+ and ASTER respectively have been compared. Although both extraction results exist some differences, the mineral type and distribution are basically the same. Both data have extracted minerals of jarosite, sericite and kaolinite with an agreement more than 78% with the distribution of known deposits, which can provide data support for prediction of metallogenic prospective area. While considering the amplitude and width of data, using ETM+ for data fusion has more advantages than ASTER.(2) Prediction results of metallogenic prospective area base on optimization algorithm of evidence weight show that the highest metallogenic probability is 0.902722 and 85.71% of known deposits fall in the four levels of metallogenic prospective area. The prediction of metallogenic prospective area in levelⅠis in good agreement with ultra-large mining area of El Salvador, which shows a satisfactory results of metallogenic prediction.(3) O n those basis, this paper further delineates three key areas for next prospecting, which are located in the western, southwestern and eastern of the study area because of the high degree of mineralization, apparent mineralization alteration anomaly and undeveloped state.