L-R restoration method of millimeter wave image based on NNSC shrinkage technique

In the restoration methods of millimeter wave (MMW) images, Lucy-Richardson (L-R) is a simple and efficient nonlinear method, but it can´t obtain better restoration result when an image is contaminated by large noise. Non-negative sparse coding shrinkage (NNSCS) technique is a self-adaptive sparse representation method and is flexible in describing data, as well as it can reduce noise efficiently without losing features of images. Considering the advantages of L-R algorithm and NNSCS method, a new image restoration method of MMW is proposed in this paper. At first, the NNSC feature bases are learned by using some clear natural images, then the noise in MMW is reduced by using the NNSCS function, and last, the restoration image can be obtained by using L-R algorithm to denoised MMW image. This modified image restoration reduces hardly the noise effect on L-R method, and improves hardly the property of L-R algorithm. Simulation experimental results show that our method proposed here is feasible and efficient in restoring image with low resolution.