Image De-Noising Based On Improved Wavelet Threshold Function

There are many image de-noising methods. In these methods, wavelet threshold de-noising algorithm is the most striking because this algorithm is simple and has a good effect. However, this algorithm also has drawbacks while has advantages above. Traditional wavelet threshold methods are divided into the hard threshold wavelet de-noising and soft threshold wavelet de-noising. The hard threshold wavelet de-noising function is not continuous at the threshold point, the image de-noised by the hard threshold de-noising method will appear pseudo Gibbs phenomenon, which will result in the distortion of the image after de-noising. The soft wavelet threshold de-noising function has the shrinkage of the wavelet coefficients, which can result in the blurring edge of a image and is not able to highlight the details of the image. In addition, the threshold for each layer of image decomposition coefficicents is fixed, it will create the coefficient of image without noise over strangled so that the effect of the de-noised image is poor.The wavelet transform is the basis of ridgelet transform. For linear singularity, the ridgelet transform can be effectively detected. Typically present singularity and line singularities in a image, so a de-noising algorithm combined wavelet threshold de-noising with ridgelet transform is even more effective. In order to make the de-noising image better, image de-noising algorithm based on wavelet transform is proposed in this paper, and ridgelet transform de-noising based on the improved wavelet threshold de-noising, the specific process is as follows:First,based on the traditional soft and hard threshold wavelet de-noising, the improved wavelet threshold de-noising function and threshold meet the following points: The improved wavelet threshold de-noising function is continuous at the threshold point so that the pseudo Gibbs phenomenon is avoided; Because the traditional threshold has "over kill" phenomenon, the improved threshold should be satisfied with the different threshold to different wavelet decomposition layers.Second,the wavelet threshold de-noising can not deal with the line singularity effectively, and the ridgelet transform can deal with the line singularity effectively. In this paper, we use the improved threshold wavelet de-noising algorithm as the basis and use the ridgelet transform method for image processing to solve the line singularity in the image effectively.Third,fuse the improved threshold wavelet de-noising image and the ridgelet transform de-noising image, so the fused image will have the advantages of threshold wavelet de-noising image and ridgelet transform de-noising image. Simulation results show that compared with the traditional soft and hard wavelet de-noising function, the improved wavelet threshold de-noising function and the ridgelet transform de-noising, the fusion of de-noising image will have a better de-noising effect.