The Highly Compressed PolSAR Model

Abstract

Additional scene information can be captured using fully polarimetric synthetic aperture radar (PolSAR) at the cost of significant increases in storage and data processing requirements. The typically sparse nature of PolSAR scenes makes compressed sensing (CS) techniques very attractive to help alleviate the increased processing and storage requirements. Here, the authors combine techniques for fast- and slow-time undersampling as well as the dropped-channel PolSAR CS technique to create a new highly compressed PolSAR model. Examples of both point target and real-world scenes are then used to demonstrate the model. Compression rates of up to 98.92% are observed for sufficiently sparse scenes.