| In this thesis, the problem is the recovery of defocus blurred image. First describe two types of commonly used defocus fuzzy model that from the traditional defocus blurred image restoration algorithm:Gaussian defocus model and pillbox model, and problems in the two models were discussed. Then study the straight edge spread function, line diffusion function and the point spread function and the Abel inverse transform function,discard the original two specific models, further improved the conventional defocus blurred image restoration method.This article is based mainly on optical theory, calculate the actual defocus blurred optical transfer function, then based on the defocus blurred circular symmetry. Then based on the circular systemstry of the defocus image,starting from the circular systemstry, Detect the straight edge spread function curve, estimate the image spread function of the image which being recovered, In this paper, the filtering is not directly using a simple Wiener filter, but its optimized, using a secondary filter, to some extent eliminates the noise.Finally the actual experiments show that the algorithms. Experiments certified the advantage of the improved algorithm, the point spread function of the algorithm is more accuracy than the traditional model. |
