Blind Restoration Of Single Motion Blurred Image Based On Variational Method

The blind deconvolution algorithm of motion blur image is an important research in the image processing field currently. In order to get the sharp image and the point spread function, the variational method is used. The variational energy equation contains the data term and regularization term. In this paper, we select the TVL1 term or TVL2 term as the data term. But the selected regularization term is the key to the variational method. So the selected regularization term need to meet the requirements of the image deblurring. The regularization term not only protect the features of the image, protect the edge and enhance the edge, but also lead the energy decrease while solving the equation and make the energy equation get its convergence much faster. In order to reduce the complexity of the solving the equation, fast method ADMM and the Split Bregman Method are introduced. The results of the experiments demonstrate the validity of the proposed method.The classical TV(Total Variation) model has been applied to the blind deconvolution of the motion blur image previously. But TV model can not be directly used to color blur image deconvolution because color image has coupling of different image layers in color images. In order to solve the blind deconvolution problem for color images effectively, we propose the CTV(Color Total Variation) model. CTV model is an extension of TV model, the model not only can blur the image, but also keep the edge of the image. Experiments demonstrate the validity of the CTV model.