Investigation of Restoration Techiques for Remote Sensing Satellite Imagery

Abstract

This paper introduces implementation of some image restoration techniques, which can be applied for satellite images. These techniques are, inverse filter, iterative method, wiener filter, regularized deconvolution filter and wavelet-based method. The restoration techniques are applied on several satellite images associated with atmospheric turbulence blur at different variance of additive noise to check the performance of each technique and its capability to restore the degraded image as close as possible to the original image. Also, comparison studies between these techniques are introduced based on measures like Peak Signal-to-Noise Ratio (PSNR) and Root Mean Square Error (RMSE). The experimental results show that: the wavelet method is the most suitable restoration technique for satellite images, since it gives high PSNR and small RMSE with respect to the other restoration techniques.