Shadow extraction of buildings from high-resolution SAR data via the combination of amplitude reversal and total variation

Abstract

For building shadow extraction from high-resolution synthetic aperture radar (SAR) data, traditional image processing methods cannot fundamentally overcome the high sidelobe interference of SAR images. In this paper, a novel technique is proposed via the combination of amplitude reversal and total variation (TV). Instead of SAR image processing, it uses SAR raw echo signal for sparse reconstruction. The technique mainly considers two scattering properties of building shadow: the zero scattering intensity and the extended block sparsity. Amplitude reversal is proposed to restrain the former, while TV is exploited to restrain the latter. Then the problem becomes a combined cost function minimization, and the gradient projection method is used for solution. Simulation results and comparison with traditional methods show the effectiveness and performance improvement of the proposed method, with fewer echo samples used.