Research on blind restoration of noisy blurred image based on deep learning
Liu Jianlan, Liu Xin, Shao Zong-jian, Rong Feng Zhong, Boyuan Ye · Proceedings of the 2020 International Conference on Aviation Safety and Information Technology · 2020
In order to solve the problem that the GMG data of the final restored image is too small and the quality of the restored image is not high, a blind restoration method of noisy blurred image based on depth learning is studied. A sufficient number of sub-pixel is used as the processing target to form the sub-pixel displacement and image registration criteria, control the displacement deviation, and construct the image denoising model by using deep learning. Under the control of the model, the non local similarity is used to process the denoised image, and the blind restoration algorithm is constructed. Finally, the blind restoration of the noisy fuzzy image is completed. In the experiment, we prepare a noisy and fuzzy image, and measure its parameters under different signal-to-noise ratio. Under the control of the parameters, two traditional methods and the blind restoration method designed in this paper are used to carry out experiments. The results show that the GMG data obtained by the proposed method is the largest, and the image quality obtained by the restoration is the highest.