Near Real Time Enhancement of Fractional SAR Imagery Based on Robust Descriptive Experiment Design Regularization Framework

Abstract:
We address new robust computationally efficient numerical technique for high-resolution reconstructive imaging as required for enhanced remote sensing (RS) with fractional synthetic aperture radar (SAR) sensor systems. The modification of the previously proposed descriptive experiment design regularization (DEDR) method via incorporating the convergence enforcing projections onto convex solution sets (POCS) is performed to enable the overall fused DEDR-POCS technique to alleviate the reconstructive problem ill-posedness and operate in (near) real computational mode. We also comment on the adaptive adjustment of the algorithmic regularization parameters directly from the actual speckle-corrupted SAR images and demonstrate the computational effectiveness of the overall proposed unified DEDR-POCS technique.