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

Image signal is often interfered by many kinds of noise when produced, transmitted and recorded. Therefore it is a very important work to use the suitable method reduction noise (Namely carries on the image noises filter) .This thesis mainly researches wavelet filter and median filter in image denoising technique.First, this article carries on the comprehensive summary to the image denoising present situation and the commonly used methods. Wavelet threshold filter method in the wavelet filter technique, which obtains widespread application because of calculated simply. After deeply researches threshold function and threshold selection in the wavelet threshold filter, this article uses the improvement threshold function and threshold to testing, the result indicates that this kind of improvement method has good denoising effect.Next, this article researches the conventional median filter, improvement median filter, maximum and minimum median filter, auto-adapted median filter. It gives a kind of median filter algorithm based on impulse detected, and studies its main performance. The simulation result shows that this algorithm has good affect to image denoising and the detail protective.Then, This article research the ways of removing mixed noises to in image. Combining median filter and wavelet filter separately in elimination impulse noise and Gauss noise merit, it gives a algorithm of removing image mixed noises. According to impulse noise and Gauss noise wavelet characteristics, it use based on the impulse noise examination median filter method to remove impulse noise and use the wavelet threshold value filter method to remove the Gauss noise. The simulation testing proof, this method has the usability and the validity, is friends with compared to the sole filter effect.Finally, this article summaries the work and makes the forecast to the algorithm of image denoising further research direction.