Automatic estimation of the noise variance in SAR images for use in speckle filtering.
Giovanni Ramponi, Raffaele D'Alvise, Cecilia R. Moloney · NSIP · 1999
Synthetic Aperture Radar (SAR) is an active imaging system widely used in remote sensing applications. SAR systems are characterized by their high image resolution and all-weather operating ability, but suffer from the notorious speckle noise [1], which is a random multiplicative phenomenon that results from coherent imaging. In this paper we propose several heuristic methods for the estimation of the speckle noise variance. They can also be applied in other cases where multiplicative noise is present, and do not need specifically tuned input parameters.