Wavelet Denoising Of Images Corrupted By Spatial Nonunifor Noise

Recent years Theory of Wavelet transformation, which is commonly accepted as one of the most splendent scientific accomplishment, leads to revolutionary influence for signal analysis image processing and other nonlinear science. Wavelet denoising is one of the important application aspects. Now, Threshold denoising which is broad adopted as its briefness.In this text, we aim at wavelet denoising method of image corrupted by non-spatially stationary additive white Gaussian noise (AWGN).On the assumption that the non-spatially stationary AWGN is slow change, partition threshold and point-wise threshold is given to single image. The latter, which is based on the neighborhood estimations using noise variances and signal variances, is self-regulating and superior to partition threshold evidently.Then, we extend united wavelet denoising method for multiple copies corrupted by spatially stationary AWGN to non-spatially case, and compares two different ways of point-wise weighted averages which are average before threshold and threshold before average. Examinations prove that the plan acquire good effect.