Application In Mineral Exploration Of Images Registration Based On Feature And Images Fusion Research

China is a big mineral resource consuming country. The proven mineral resource will not meet the need of economic development in next twenty years. Using Remote Sensing to prospect is fast and efficient, so there are many scholars who value it.This paper takes Zijingshan Pingshuo Shanxi Province as the study area and chooses optical SPOT image and SAR image as the data source. It Conducts the following research:(1) It preprocesses the SPOT image and SAR image. In order to fully reserve the texture informations of the SAR image, it uses E(I), ENL and edge retention extent to analyze and compare Frost, Lee, EnhanceLee, Mean filters and so on, then chooses the better Lee filter for the next experiments.(2)The registration between optical image and SAR image. This paper analyzes the advantages and disadvantages of the traditional SIFT (Scale Invariant Feature Transform) registration method. Using the local feature points match method modifies the traditional one. The modified SIFT method makes the optical image and SAR image registration improved greatly in accuracy, precision, timeliness and other aspects.(3) The fusion of the images. It analyzes the color and spectral distortion of HIS, PC (principal component transform) and Brovey fusion about the optical image and SAR image. In order to overcome these drawbacks, it modifies the traditional Brovey fusion, and then judges them by the degree of spectral distortion and other standards. The modified fusion method is the best one.(4)The extraction and analysis of prospect informations. It uses different methods to extract linear and cyclic informations, and then interpret them. As for the vegetation abnormal informations, it adopts classification. It conducts a buffer analysis and overlay analysis to reduce the fortuity of single information and improve the accuracy and reliability of prospect areas.