A new approach for one-bit compressed synthetic aperture radar imaging

Synthetic Aperture Radar (SAR) imaging system requires too much measurements to achieve a high-resolution image. This situation also raises the need for fast analogue-to-digital-converters and large on board storage systems due to the large number of measurements. Using 1-bit quantization may be a rational solution to these issues. A new framework is introduced for 1-bit compressed SAR imaging by using time-varying thresholds in this dissertation. It has been demonstrated how to reconstruct sparse SAR images from noisy measurements which are quantized to 1-bit with time-varying thresholds. In conventional 1-bit Compressive Sensing (CS) algorithms 1-bit quantization has been done by comparing measurements to a zero threshold. This makes the magnitude information of the signal to be lost and exact signal recovery becomes impossible. Unlike those algorithms, we do 1-bit quantization by comparing the received signal to time-varying thresholds. With this 1-bit quantization method, unit-norm constraint, consistency function and sophisticated optimization algorithms are no longer needed. Moreover, the amplitudes of the signals obtained by this method are more accurate. Using this 1-bit quantization approach, we can formulate 1-bit CS SAR imaging reconstruction problem as an unconstrained optimization problem where the objective function includes l_2 data-fidelity term and a non-smooth regularization function. For solving this unconstrained optimization problem, we use variable splitting and the alternating direction method of multipliers which is computationally efficient and easy to implement. The results from simulations with real and synthetic SAR images validate the effectiveness of the proposed algorithm named as BCST-SAR (Binary CS with Time-varying thresholds in SAR imaging).