Log-Likelihood Method of Reducing Noise in CRISM Along-Track Oversampled Hyperspectral Images
Christina D. Kreisch, R. E. Arvidson, Joseph A. O’Sullivan, Ke Li, Daniel Politte, Justin Finkel, E. A. Guinness, N. Stein, Abigail A. Fraeman · 2015
CRISM hyperspectral measurements are modeled accounting for spatial and spectral pointspread functions. Surface hyperspectral image recovery is ill-posed. Penalized maximum likelihood estimates yield reduced noise in each spectrum estimate, allowing identification of subtle spectral absorptions.