Blind Image Restoration Based On Variation Method And LPA

In image acquisition, transmission, and record keeping process, due to various factors, the degradation of image quality inevitably arise. The main performances of the image degradation are blur and additional noise. In order to recover the original clear image from the degraded image, it is necessary to use image restoration techniques. The point spread function and noise distribution in the circumstances is known in traditional image restoration approach. But in practical applications, we often do not know the prior knowledge; we need only to restore degraded images to get the original image, which is to be blind restoration problem. The purpose of this paper is to achieve the blind restoration of noisy and blurred images.The additive noise in blurred image will affect the quality of the restored image, the paper studies the recursive LPA-ICI denoising algorithm, it can remove the noise and greatly save the image details. After that, we use two blind image restoration algorithms, single channel blind restoration algorithm and multi-channel blind restoration algorithm. In the single channel blind restoration algorithm, we solve blind image restoration problem with two parts. First, to eliminate noise and enhance the edge information, we use the method of combining Canny edge detection and wavelet thresholding denoising; Secondly, construct a suitable energy regularization function using the image inherent prior knowledge, ultimately obtained a clear original image by seeking the minimum value. In the multi-channel blind restoration algorithm, we use frequency domain gradient projection method, the image restoration and blur identification are processed in the same time, the algorithm is relatively simple to implement, in the iteration process of the algorithm, use recursive LPA-ICI denoising for images and point spread function to reduce the impact of noise on the algorithm.Blind restoration of noisy and blurred images is a very complex issue, the two recovery algorithms used in this paper has a strong anti-noise performance, and image restoration results are very good. While propose new algorithm, the combination of image denoising and image restoration is an effective way to improve the quality of the restored image.