We present the problem of simultaneously increasing the spatial resolution and deconvolving channels of hyperspectral images where the blurring kernels are unknown. A high resolution image is incorporated into a directional total variation prior for the corresponding variational model. The non-smoothness and non-convexity of the objective function is treated using the PALM-algorithm. Numerical results on remote sensing data show the potential of the proposed method and suggest that it is robust with respect to mis-registration.

This presentation is part of Minisymposium “MS55 - Advances of regularization techniques in iterative reconstruction (2 parts)”
organized by: Zichao (Wendy) Di (Argonne National Lab) , Marc Aurèle Gilles (Cornell University) .

Authors:

Leon Bungert (University of Münster)

Matthias J. Ehrhardt (University of Cambridge)

Carola-Bibiane Schönlieb (University of Cambridge)

Marc Aurèle Gilles (Cornell University)

David Coomes ( Forest Ecology and Conservation Group, Department of Plant Sciences, University of Cambridge)

Rafael Reisenhofer (University of Bremen)

Jennifer Rasch (Fraunhofer Heinrich Hertz Institute)

Keywords:

hyperspectral imaging, image deblurring, image fusion, inverse problems