Using deconvolution and machine learning to improve MSL Curiosity images: application to ChemCam and MAHLI

Stéphane Le Mouélic (1) , O. Gasnault (2) , K. Herkenhoff (3) , W. Rapin (4) , K S Edgett (5) , R A Yingst (6) , Nicolas Mangold (1) , Gwénaël Caravaca (1) , P. Pinet (2) , R. Wiens (7) , S. Maurice (2)

1 LPG - Laboratoire de Planétologie et Géodynamique [UMR 6112]
2 IRAP - Institut de recherche en astrophysique et planétologie
3 Astrogeology Science Center [Flagstaff]
4 Division of Geological and Planetary Sciences [Pasadena]
5 MSSS - Malin Space Science Systems
6 PSI - Planetary Science Institute [Tucson]
7 LANL - Los Alamos National Laboratory

We investigate the performance of the recently released image processing software suite “Topaz Gigapixel” on martian images from the Mars Science Laboratory “Curiosity” rover. The objective is to evaluate the possibility to visually enhance the apparent resolution of acquired images, using cutting-edge image processing algorithms, in order to improve the mapping and interpretation of small scale features such as layers, laminations, cross-bedding, grains, veins, nodules...

Domaines: Sciences de la Terre; Planétologie; Optique [physics.optics]; Machine Learning [stat.ML]

Stéphane Le Mouélic, O. Gasnault, K. Herkenhoff, W. Rapin, K S Edgett, et al.. Using deconvolution and machine learning to improve MSL Curiosity images: application to ChemCam and MAHLI. 51st Lunar and Planetary Science Conference, Lunar & Planetary Institute, Mar 2020, The Woodlands, Texas, United States. ⟨hal-02526554⟩