L1-norm Regularization Based SAR Phase Error Estimation with Incomplete Data

Abstract:
The received data of synthetic aperture radar (SAR) will be incomplete or corrupted for many reasons, e.g.,contamination of time series and interference of other transmitters operating in adjacent bands. Therefore, pre-processing of the received data such as phase error compensation needs to be performed. However, classical autofocusing algorithms are hardly applicable to incomplete echo data. In order to solve this problem, this paper proposes a novel method of phase error estimation with incomplete data which is versatile and not restricted by the type of phase error and SAR imaging mode. Firstly, by taking the phase error into account in the model construction, the proposed method estimates the phase error by solving an L1-norm regularization problem. Then BiIST algorithm is used to obtain the phase error by iterative operation. Experimental results based on typical spaceborne SAR parameters show that the proposed method can achieve the accurate estimation of phase error and the compensated echo can be directly used for subsequent tasks such as imaging, target detection, and multilooking.