Abstract:In order to solve the limitations of existing depth restoration methods, an image depth restoration method based on Moreau envelope and iterative re-weighting strategy is proposed. A non convex penalty function based on Moreau envelope is proposed, which improves the prior sparsity of the model and keeps the convexity of the model. The convergence of the algorithm is analyzed. Then an iterative re-weighting algorithm is introduced to deal with the depth color inconsistency problem. In addition, an accelerated algorithm is proposed to transform the depth observation matrix into Fourier domain for fast processing. Finally, the experimental results of data sets show that the method can deal with various types of deep degradation, and has achieved good results in terms of recovery accuracy and running time.