Image Denoising Based On Wavelet Transform And Its Application In SAR Image

Image is an important information source for human beings. However, digital image is usually corrupted by the noise in its acquisition or transmission. It leads to the image quality deline, which is very unfavorable to higher level image processing. There is a dilemma contradictions between edge information and noise suppression protection in the traditional method of image denoising. Therefore, people have been pursuing a method, which can denoise effectively and keep the edge information simultaneously. The wavelet theory praised as'mathematical microscope'has been rapidly developed. For the characteristics of its low entropy, multi-anaylsis and relativity, the wavelet transform has been widely applied in the field of denoising. The wavelet theory can not only wipe off noise but also retain the image details. Therefore, the wavelet transform for image denoising algorithm has practical significance.Based on the profound analysis on wavelet image denoising, three classical wavelet denoising methods, which is wavelet transform modulus maxima denoising method, the relativity of the wavelet coefficient denoising method, wavelet threshold denoising method are introduced in detail. The advantages and disadvantage of these methods and their applicable condition are discussed at last and the simulation experiments show the results of image denoising. The conclusion is that wavelet threshold denoising is the best way and calculate is minimum.Discontinuity of the hard-thresholding function results in Pseudo-Gibbs phenomnon of the reconstructed signal. Soft-thresholding function has good continuity but a constant deviation of the estimated value from the actual value confines its application. To improve the performance of wavelet threshold denoising algorithn, the thesis also proposed a threshold improved algorithm for existing algorithms, the new threshold denoising method based on improved threshold function and noise variance estimation. The method chooses a new threshold function with multiorder continuous derivatives, which makes it possible to construct an optimal threshold through the method of iteration operation. At last, the result of the simulation experiments show that the new method usually obtion better methods. The new method can also preserve more image details. The image of the peak value signal-noise ration and visual effect got obvious improvement, so it is an effective method on image denoising.At the end of thesis, the imporved algorithm was applied to synthetic aperture radar(SAR) image which was interferenced by coherent speckle noise. The experiment results illuminates that the method of this thesis can preserves the SAR image structure and may suppress the speckle noise. Compared with traditional methods, the improved method has better denoising.