SAR speckle filtering and automatic feature extraction: An innovative integrated algorithm

Filtering processes are necessary in order to extract automatically useful information from Synthetic Aperture Radar (SAR) data. Galileian Plus S r.l. (GPLUS) developed an innovative proprietary procedure (European Patent EP2146315 of 2009), which takes off from the demand of maintaining high spatial details of nowadays SAR sensors during the filtering process, while at the same time improving the data radiometric resolution. In this article, the innovative filtering procedure is described in details, by giving particular importance to mathematical definition of the non-linear transfer function G_α,β. Comparisons with the state-of-the-art of speckle filters show the enhancement and benefits of GPLUS filter, not only for visual interpretation of the data but mainly for automatic classification and feature extraction purposes. The two applications described in the paper show the potentiality and the step ahead performed with this algorithm: automated classification and extraction processes based on SAR imagery move straight towards a wider range of applications.