The Algorithm Research Of Image Denoising Based On Median Filter And Wavelet Transform

Nowadays society has entered the new era of digital information, image signal hasbecome a very important media and way to attain and transport the information. Image signalis often interfered due to disturbance from different kinds of sudden noise in the process ofgeneration, transmission and recording, which make a bad the impact on the following imageprocessing. For instance, image segmentation, feature extraction, etc. Therefore, before thefollowing processing, we should process the image to reduce the noise first, which is an veryimportant task.This article mainly gives an account of the most common image noise, impulse noise andGaussian noise, using the image denoising technique of wavelet filter and median filter. Thisresearch contents consist of the following three aspects:First, this article carries on the standard median filter, improvement median filter andextremum median filter algorithms. It gives a kind of improvement median filter algorithmand studies its main performance after analysing the merits and dismerits of differentalgorithms. Comparing it with other algorithm by simulation experiments, the result showsthat this algorithm has good affect to the detail protective and image denoising．Next, this article does some researches on the image denoising methods based onwavelet. After analyzing wavelet coefficient characteristic of natural images, it deeplyresearches threshold function and threshold selection in the wavelet threshold filter, it usesbased on the impulse noise examination median filter method to remove impulse noise anduses the wavelet threshold value filter method to remove the Gauss noise. The experimentalresults indicate that the algorithm is better than other wavelet denoising algorithms whenremoving Gaussian white noise．Finally, this article focus on that one denoising method is only effective to some one kindof noise in common cases．In order to suppress the mixed noise，this paper gives an image denoising algorithm, which combines the adaptive median filter based on extremum medianfilter with the wavelet denoising based on Gaussian Mixture Model．The simulation testingdemonstrates that this method has the usability and the validity compared with the sole filtereffect.