An Exploratory Analysis of Speckle Noise Removal Methods for Satellite Images

Satellite images captured in a variety of modalities serve as the primary source for many applications. Satellite image processing extracts the image /spectral information represented in the form of pixels, classifies those pixels based on the similarity measures and further analyzes the inherent data, as per the requirements. The foremost objective of satellite processing is to automatically categorize the pixels in an image into the respective land cover class labels or themes. These pixels are classified by its spectral information and it is determined by the relative reflectance in various bands of wavelength. The accuracy and outcomes of any satellite image processing procedure, irrespective of the application domain, directly depends on its quality. Satellite images are invariably degraded by speckle noise. Hence, preprocessing the images for speckle noise suppression and/or cloud removal is deemed an inevitable component in satellite image processing. Researchers have proposed a spectrum of methods for speckle noise/cloud removal. A detailed review on the significant research publications on speckle noise removal are summarized in this article. The consolidation of methodology merits and demerits of the select research articles are presented in this paper. This review article on speckle noise removal is designed as a ready-reference for those researchers working in satellite image processing.