Application Of Wavelet Transform To Image Processing

Wavelet analysis theory, as a new time-frequency analysis tool, has been well applied in the area of signal analysis and processing. An image is actually a two-dimensional signal. So it is natural to apply wavelet analysis to the area of image processing. Image de-noising and edge detection are two widely used technologies in image pre-processing. By enhancing SNR and highlighting expected features of image, it will be more convenient for further step of processing. Wavelet transform is more and more frequently applied to image processing according to its own advantages.First, beginning with basic theories, wavelet transform is thoroughly and deeply introduced in this article. A series of related contents from definition to wavelet analysis are gradually elaborated step by step. Mallat algorithm which is of great importance in wavelet analysis is introduced in the end.Second, researches on application of wavelet transform to the areas of image de-noising and edge detection are the major part of this article. For comparison, several traditional image de-noising methods and edge detection methods are briefly introduced. The theories and steps of completion of image de-noising and edge detection algorithms to which wavelet transform is applied are elaborated. Three different threshold functions include setting wavelet coefficients of high frequency zero, hard threshold and soft threshold are separately applied to image de-noising. Using SNR and PSNR as evaluation criterion, image de-noising and edge detection algorithms between wavelet transform and traditional methods are simulated and compared to draw a conclusion.