Classification technique for lunar geologic mapping

Abstract

Goal of this study is a proposal of algorithm to help geologic mapping on the basis of massive data derived from the above-stated SELENE mission. A classification technique by clustering which takes band luminance as feature quantity has been adopted for a geological classification in the past; however proper results have not been presented because band luminance is affected by the topographical effects and capturing timing. Therefore various texture features extracted from band images are also provided as the feature quantity, and the result of study if these bring a better result will be presented.