Analysis of SAR images speckle reduction techniques

Remote Sensing is the process to acquire information about phenomena or an object without physical contact with the subject under consideration. The applications of remote sensing vary from Geography to land surveying and many more. Optical sensors used reflected sun rays for object identification, being usable in day light and shorter wavelength makes it inefficient for limited time availability. Thus it cannot penetrate the earths' surface and is affected by atmospheric factors as well. Radar based remote sensing are now used to overcome the shortcomings of optical sensors. Images acquired by receiving the transmitted signals in radar sensors encounter speckle noise phenomena which is a granular noise. The noise exists inherently in SAR images due to interference of received signals either due to constructive or destructive addition of received signals that degrades the image quality and thus does not give accurate information. To denoise the images several techniques have been proposed and applied which includes local Filters, non-local filters and transforms. Each technique has its merits and demerits. Recent focus has been shifted towards the combined solution of non-local filters with transforms. The motivation of this study is to study the existing hybrid solution and work towards new venues in hybrid techniques to reduce information loss and produce better results.