Estimation of Lunar FeO and TiO₂ with Support Vector Regression Analysis and Evaluation of the Stokes Parameter Using M3 and LRO Mini RF Data

Abstract. Moon Mineralogy Mapper (M3), the hyperspectral sensor of ISRO’s Chandrayaan 1 mission, dedicated to map the surface mineral composition, provided an opportunity to map the lunar regolith in Global and Target modes. These high resolution hyperspectral data is used to recalibrate the pioneer work of Lucey et al. (2000) and thus estimated the FeO and TiO₂ contents of the regolith. As the FeO and TiO₂ estimations have to handle a huge amount of data, Support Vector Regression analysis (SVR) is introduced to reform the FeO and TiO₂ wt % equations for Apollo and Luna landing sites. These equations with the optimized origin are used to estimate the FeO and TiO₂ contents of the target locations. Catharina crater of the lunar near side is taken as the study area and FeO and TiO₂ wt% of the crater are estimated with M3 data. Composite chemical content (FeO wt%+ TiO₂ wt %) of the Catharina crater is estimated. LRO mini RF, Hybrid-polarized, dual-frequency synthetic aperture radar of LRO mission is used to characterize the back scattering properties of the crater surface. Stokes parameters are extracted from the LRO mini RF SAR data for the same study area. Composite chemical content estimated using M3 data is compared with the Stokes total intensity parameter extracted from the LRO mini RF data. The total intensity parameter (Stokes vector S1) is directly proportional to the composite chemical content and thus shows a linear relationship.