Blind deconvolution of atmospherically degraded infrared images using normalized sparsity measure

Blind image deblurring from a single image is a highly ill-posed problem. To tackle this problem, prior knowledge about the point spread function (PSF) and latent image are required. In this paper, a blind image deblurring approach is proposed to remove atmospheric blur, which utilizes the normalized sparse prior on the latent image and radial symmetric constraint on PSF. By introducing an expanding operator, the original constrained minimization problem is simplified to an unconstrained minimization problem and it therefore can be solved efficiently. Experiments on both synthetic and real data demonstrate the effectiveness of our approach.