A regularization LR-algorithm for restoring images on Gaussian noises model
Weihao Liu, Xuan Cai, Lin Mei · 2012
Ordinary Lucy-Richardson (LR) restoration algorithms are used to restore high SNR degraded images including astronomical images and achieve good results. The algorithms are very sensitive to noises and use the assumption—noises observe the Possion distribution. However, there are always Gaussian noises in natural images. In this paper, we propose a regularization LR-algorithm based on the Gaussian noises model. Our algorithm involves two regularization methods. One gives regularization restriction to the residual signal of the restored image. The other gives regularization restriction to the sparse gradient of the restored image, which is different with the conventional gratitude restriction. They can efficiently suppress the amplification of noises and preserve the details of images. Finally, we show the advancement of our algorithm using some experiment data.