Satellite Image De-Noising With Harris Hawks Meta Heuristic Optimization Algorithm and Improved Adaptive Generalized Gaussian Distribution Threshold Function

Abstract

An image may be influenced by noise during capturing and transmitting process. Removing the possible noise from the image has always been a challenging issue due to this fact that further processing will not be possible unless by diminishing the noise from images. Many researchers attempted to remove the noise to improve the qualitative and also the quantitative results but these methods could not preserve the quality of images after applying de-noising techniques. In this paper, in the first stage, we utilized the most recent nature-inspired meta-heuristic optimization algorithm to get the optimal solutions for the parameters of thresholding function. Using the Harris hawk optimization (HHO) algorithm results in obtaining the optimized thresholded wavelet coefficients before applying the inverse wavelet transform. In the second stage, we proposed the improved adaptive generalized Gaussian distribution (AGGD) threshold, which is a data-driven function with an adaptive threshold value. This function can be fitted to any kind of images without using any shape tuning parameter. It is clear that the calculation of the threshold value does not require any optimization and LMS learning algorithm. The qualitative and quantitative results validate the superiority of the proposed method.