An Improved Restoration Algorithm for Blurred Images Based on Complete Blind Convolution and Non-local Means Filtering

A method of estimating point spread function is designed in this paper. This PSF is applied in complete blind deconvolution for the restoration of defocus blurred images. The ringing artifact which is produced in the procession of image restoration needs to be suppressed. By using histogram to judge the type of random noise in the image, non local means filtering is chosen to denoise. Considering the long running time of the NLM filtering, only the wavelet low frequency coefficients are denoised with it. MATLAB simulations show that the designed PSF and this method have better effect of restoration and shorter running time.