Conditional Chromatic Filtering for Restoring Pansharpened Images

Pansharpening techniques aim at fusing the geometry of the panchromatic and the spectral accuracy of the low-resolution multispectral data provided by Earth observation satellites to produce a high-resolution multispectral image. Despite significant progress in the field, spectral and spatial distortions might still compromise the quality of the results. We introduce a restoration strategy to mitigate artifacts of fused products. After structure-chromaticity decorrelation of the pansharpened image, the chromatic components are filtered conditionally to the geometry of the panchromatic by means of a convex nonlocal variational model. The structural component is then replaced by the locally histogram-matched panchromatic for spatial enhancement. Experimental results illustrate the efficiency of the proposed restoration chain.