On Selection of the Observation Model for Multilook Application of Sparse Microwave Imaging

In this paper, we compare the performance of multilook application for sparse microwave imaging with two different linear models. One is based on the inverse of conventional multilook procedures with frequency-domain multilook observation (FMO) and the other employs the decomposition of synthetic aperture to yield a time-domain multilook observation (TMO). From the comparisosn, we found that FMO is more stable and can achieve better speckle reduction in the multilook application. In addition, benefited by FFTs type operations, the FMO can also be implemented much faster than TMO. These facts then suggest the using of FMO in multilook sparse microwave imaging.