The Research On Remote Sensing And Multi-Information Fusion In Baoxingchang Copper Mining Area

Fusion is an important information processing tool.Since 1960s,the information fusion develops a very complete theoretical system due to scholar's efforts.The information fusion is divided into pixel level,feature level and decision level according to its architecture.However,the previous remote sensing and mutil-information fusion research is more limited to the single remote sensing data,and focuses on the research of pixel level rather than feature level and decision level.The remote sensing and multi-information that composed by geophysical,geochemical exploration occupy a pivotal position because of its multi-methods and data sources in the exploration of mineral resources.In order to make up for the shortcomings of the past research,this thesis selected Baoxingchang copper mining as the research area,using remote sensing,geophysics,geochemical data,based on IHS,principal component transformation and wavelet transform and other methods to carry out pixel-level remote sensing information fusion research.With geochemical data,tectonic density,ring structure,mineralization alteration zoning information,this thesis constructs remote sensing multi-information fusion model of C4.5 decision tree to explore the establishment of an optimized fusion method.The results of demonstration applications with remote sensing multi-information fusion experiments in Baoxingchang copper mining area detected the metallogenic and ore-controlling effects of the copper deposit.This thesis's major achievements are as follows:(1)In pixel level,the comparative study of remote sensing multi-information fusion experiment based on IHS,principal component transformation and wavelet transform,is carried out,and the multi-information model of wavelet transform remote sensing is optimized.The wavelet transform fusion research with GeoEye-1 remote sensing image and geophysical and geochemical data,not only can preserve the original hue information of the original remote sensing image,and also highlight the geophysical and geochemical anomaly information.PCA fusion is a better among the remote sensing and multi-information fusions,in which PC4 weighting fusion with remote sensing image and geophysical and geochemical data is much better than the other.Although the original image hue information is better preservedin PCA fusion,the effect is not well better than the wavelet transform fusion.(2)In decision level,with the remote sensing geological interpretation,geochemical element cluster analysis and geophysical,geochemical data interpolation,the remote sensing line,ring structure,alteration anomaly and Cu and related copper mineralization abnorming information is carried out.On remote sensing multi-information fusion method of C4.5 decision tree,the result of the experimental analysis and field verification show that the decision tree fusion can reflect the minenalization more clearly,which can better reveal the relationship between remote sensing multi-information then refecting ore and ore control features.By this way,the inclues is better supporting Baoxingchang Copper mining area and similar areas of copper exploration.