Compressive Acquisition - Sparse Representations and Compressive Sensing for Imaging and Vision - page 24

Fig. 3.5 512 × 512 Shepp-Logan Phantom example. (a) Fourier domain sampling pattern.

(b) Back-projection. (c) L1TV reconstruction. (d) GradientRec-LS reconstruction

three to five times the number of sparse coefficients [20], [79]. This means that
GradientRec-LS can recover gradients perfectly from 13107

= . × ×

0 05 512 512

. × . × .

compressive measurements, which is approximately 6 15 X y 0 and 4 4 X x

0

=
17544 measurements. Hence, 13107 measurements are not enough for the TV-
minimization to recover the underlying sparse gradient image.

×

Whereas for a good reconstruction, TV-minimization requires about 4 4386

3.3 Compressive Synthetic Aperture Radar Imaging

Synthetic aperture radar is a radar imaging technology that is capable of producing
high resolution images of the stationary surface targets and terrain. There are four
common modes of SAR: scan, stripmap, spotlight and inverse SAR (ISAR). In this
chapter, we will mainly focus on the spotlight mode SAR and ISAR. In spotlight
mode SAR, the radar sensor steers its antenna beam to continuously illuminate the