Image Denoising Using The Minmum Energy Based On The Maximum Entropy And It's Improves

Image is obtained by image acquisition equipments such as cameras, which can reflect the informations of the real scene. No matter optics, optical, or electronic methods will the images be degraded. There is an important problem in digital image processing to solve,which is how to get the image from the known data to reflect the reality of the objective scene.Image denoising plays a very important role in image processing. With the development of the science and technology, the image quality is required more and more high. The noises make the image recognition rate reduced, and important information lost.At the same time, it can let the human's visual uncomportable.. Effective image denoising algorithm can make full use of prior knowledge of the image, establish the mathematical model of image denoising, and then through the corresponding algorithm processing, to get clear of the original image. Image denoising technology can remove or reduce the problems happened in the process of digital image, so as to make the image as much as possible close to the real scene.In this dissertation, based on the maximum entropy principle and the minimum energy method, I proposed a model for image denoising taking full advantage of image information. The model will make full use of the noised image,so as to maximize the image of the original information. Based on Matlab and OpenCV development tools, digital image processing knowledge, probability theory and other disciplines, a general significance denoising algorithm is designed. The main work can be divided into the following several aspects:1. Propose the maximum entropy principle and the minimum energy method of image denoisingmodel.In the process of image how to get to know the unknown information by denoising, can be thought of as a probability problem. How to maximize the probability is considered in this dissertation the main point.2. To improve the denoising model, makes the algorithm not only has better denoising effect to the low variance gaussian noise, at the same time for high variance gaussian noises, impulse noise and gaussian pulse noise have better effect.3. According to the above model proposed, experiment analysis and comparison with other better algorithms, we summarize the merits of the algorithm in this dissertation.In this dissertation, all the algorithms have been programmed. Now it is on a large scale of the experimental stage...