Independent Component Analysis of the Gamma Ray Spectrometer data of SELENE (Kaguya)

Abstract

We analyze the spectra measured by the Gamma Ray Spectrometer (GRS) on board the SELENE satellite orbiting the Moon. The spectra consist in 8192 energy channels ranging from 0 to 12 MeV and exhibiting lines of interest (O, Mg, Al, Si, Ti, Ca, Fe, K, Th, and U) superposed on a continuum. We have also analysed the data with various multivariate techniques, one of them being the Independent Component Analysis. We have used the JADE algorithm for our analysis that we focused in the energy range from 750 to 3000 keV. We identify at least three meaningful components. The first one is correlated to the Thorium map. The corresponding correlation coefficient spectrum exhibits the lines of Thorium, Potassium and Uranium. The second component is clearly correlated with the Iron as shown on its corresponding spectrum. Finally the third component seems to be related to the altitude of the spacecraft. This work shows that maps of elements such as iron will be available with the GRS data by a purely statistical analysis.