Research Of Image De-noising Based On Ridgelet Transform

Many images always have noise. In order to further image analysis and communication,the noise needs to be reduced in image preprocessing. The traditional methods can filter noise,but at the same time they make the image details blurred. Wavelet transform has the characteristic of"mathematics microscope", thus it can not only wipe off noise but also retain the image details. Ridgelet transform is a new kind of multi- scale analysis technique after wavelet transform. For image processing, ridgelet transform is more effective than the wavelet transform in representing linear and super-plane singularities.In the thesis, we study the application of wavelet and ridgelet transform in image de-noising and some threshold based de-noising methods, such as DJ, BayesShrink, NormalShrink etc., then propose and implement two algorithm of image de-noising. The first algorithm improves the Lakhwinder Kaur's NormalShrink based on the idea: noise is decreasing while decomposition level is increasing. The second algorithm proposes a stationary ridgelet transform, which is obtained by substituting the stationary wavelet transform for orthogonal wavelet transform. The stationary wavelet is a new transform method based on wavelet transform, which can reduce noise effectively with good image edge character. The stationary wavelet de-noising has obvious superiority compared with orthogonal wavelet de-noising. Finally, these algorithms verified the validity through simulate experiments.