Abstract:In areas where geological work is relatively low and bedrock is well exposed, the use of remote sensing data should be paid enough attention. Thus, the working target area can be effectively reduced and the efficiency of prospecting can be improved. In this paper, taking Xishui area in Gansu province as the study area, based on ETM+ remote sensing data, by using multivariate data analysis, normalization, mask analysis, principal component analysis, density segmentation method, extraction of gold deposit from Landsat7 ETM+ data information. On the basis of remote sensing geological interpretation, the extracted hydroxyl anomalies and iron stain anomalies have been fully analyzed. Combining with regional geological characteristics, regional metallogenic conditions and the characteristics of regional geophysical and geochemical anomalies, the fake and weak anomalies can be removed as far as possible. It is showed that this method can effectively eliminate the influence of disturbance information, such as vegetation, water system and shadow, and extract faint gold mineralization alteration information from remote sensing data carrier. The mineralized alteration information extracted has a good correlation with the fracture structures, and mainly distributed in the metallogenic dominant position of fault structure intersection, fracture and rock mass intersection, which has a certain guiding significance for prospecting work in the future.