The Investigation Of The Digital Image Restoration Techniques

Image is human visual base, and it gives us idiographic and intuitionistic sense. Image digital mode involves two steps, sampling and converting to quantity. Digital image processing is to transform the image signal to digital format, and then use the computer to process.Digital image processing techniques involve many aspects, and two aspects are discussed in this dissertation, which are image zooming and image restoration. At first, we introduce the background, the applications, the development conditions of the digital image processing techniques and the main contents, then particularly introduce three interpolation methods in common use at present. And furthermore, the interpolation method by vector valued continued fractions is introduced. A new method for image zooming is presented, which comes into being by combining the linear method based on neighborhood pixels and the nonlinear method based on Newton-Thiele's vector valued interpolation. And plenty of experimental results show that the method is effective.And then, we introduce the reason resulting in image degeneration, the category of image restoration technique, several image restoration methods in common use at present, and some other image restoration methods. A new improved method for image restoration is presented in this dissertation, which is based on least-squares method, regularization method and steepest descent algorithm. The fourth-order partial differential equations method and the least-squares increment iterative and regularization method are combined to recover the degraded image, and the experimental results demonstrate the feasibility and validity of the methods.