Research On Improved Image Denoising Methods Based On Wavelet And Contourlet Transform

In the process of collection, processing or transmission, noise is involved in, anddebases the quality of digital images. Noise is not conducive to subsequent imageanalysis, so image denoising is essential for image processing. By analysing thetraditional image denoising methods based on wavelet and contourlet transform,improved image denoising methods are proposed. The paper mainly involves thefollowing:(1) Elaborate the background and significance of image denoising and analysedevelopment status and trends at home and abroad of image denoising. Then introducethe theories of knowledge of wavelet transform and contourlet transform and traditionalimage denoising methods based on wavelet transform and contourlet transform. Theseprovide theoretical basis for the subsequent improved image denoising methods.(2) Image denoising based on wavelet transform is effective to Gaussian noise, butpoor to impulse noise. To remove impulse noise, Gaussian noise and impulse-Gaussianmixed noise effectively, image denoising based on second derivative operator andwavelet transform is proposed. First of all, we construct an image denoising methodbased on second derivative operator to noise location, which can denoise impulse noiseeffectively. Then, the threshold of the traditional wavelet threshold denoising method iscorrected. Finally, image denoising method based on second derivative operator to noiselocation and the improved wavelet threshold denoising method are combined.(3) Traditional infrared image denoising based on wavelet transform is usuallydivided into the thresholding denoising and the inter-scale correlations denoising, and isnot effective to infrared image. Through the analysis of their defects, infrared imagedenoising based on improved threshold and inter-scale correlations of wavelet transformis proposed, which improves comprehensive ability to remove noise on the infraredimage, with good usability.(4) Considering the limitations of wavelet transform and thresholding functions,uniform thresholding function named YW thresholding function based on contourlettransform is constructed. YW thresholding function takes into account the influence ofnoise, image and contourlet, has self-resilience and universality, and can carry out betterthreshold processing in each case.