Restoration Method for Satellite Image Based on Content-Aware Reciprocal Cell Pool

摘要

The challenge for the VHR satellite image restoration is how to cope with the simultaneous deblurring and denoising problem. Although the adaptive reciprocal cell (AR-cell) can be used to suppress noise in image restoration, it is an isotropic linear filter. In this letter, we analyze the AR-cell model, and then a piecewise linear version is developed to adapt to different similar image structures, giving rise to an AR-cell bank. Lastly, based on the bank, a group-wise regularization model is introduced for image restoration. Experimental results demonstrate the promising performance of the proposed method.