Research On Image Restoration Method Integrating MMTD

Image restoration is a process of reconstructing an image from degradation and is an importantresearch topic in the field of digital image processing. Since noise is one of the main causes ofimage degradation, it is an important research topic in the field of image restoration as well as themain research content of this paper on how to remove the noise.This paper analyzes the reason of image degradation and the degradation model and studies thetheory of medium mathematical systems and wavelet transformation, based on which threeimproved image filtering method based on Medium Truth Degree and Wavelet transformation areproposed. The main contents of this paper are as follows:（1）This paper studies a medium filtering algorithm and proposes a new algorithm combinesmedium filtering algorithm and median filtering algorithm in order to solve the original algorithm’sproblem of residual noise point and slow speed when removing salt and pepper noise. The newalgorithm combines the effectiveness of median filtering algorithm and the efficiency of mediumfiltering algorithm and thus has better filtering effect and faster processing speed compared withmedian filtering and medium filtering algorithm when removing salt and pepper noise.（2）This paper studies the principle of medium filtering algorithm on Gaussian noise andproposes an improved medium denoising algorithm based on the principle and distributioncharacteristics of Gaussian noise. The new algorithm improves the delineated rules of mediumpredicate area and the mathematical methods denoising Gaussian noise. Experimental results showthat the new algorithm has significant improvements compared with medium filtering algorithmespecially when the intensity of Gaussian noise is high.（3）This paper studies the wavelet domain image denoising algorithm and proposes animproved hard threshold VisuShrink denoising algorithm based on detailed analysis of hardthreshold denoising algorithms. The new algorithm smoothes the image in the process of hardthreshold denoising and thus overcomes the ringing effect of hard threshold denoising method.Experimental results show that the new algorithm keeps more image detail and contrast and thushas a better noise immunity.