Paper Title
SATELLITE IMAGE DENOISING USING LINEAR WAVELETS

Abstract
Modeling of two major linear wavelet denoising techniques; Poisson Unbiased Risk Estimate Linear Expansion of Thresholds (PURELET) and Stein’s Unbiased Risk Estimate Linear Expansion of Thresholds (SURELET) in application to satellite images is presented in this paper. Satellite image sensors are often corrupted by noise that is effectively enhanced by estimation of noise and robust denoising algorithm application. By dynamic thresholding of unbiased estimates, optimized enhancement of spatial resolution can be obtained. Poisson unbiased risk estimate (PURE) is an unbiased estimator of the mean square error between the original and estimated images. Stein’s unbiased risk estimate (SURE) is the apriori estimation of Mean Square Error resulting from an arbitrary processing of noisy data for the additive Gaussian noise model. By minimizing the unbiased estimates using Linear Expansion of Threshold (LET), noise can be significantly reduced and resolution can be enhanced.