Particle Swarm Optimization를 이용한 다중편파보정 성능향상 연구

Abstract

With the development of state-of-the-art low earth orbit observation satellites, polarimetric synthetic aperture radar(SAR) images have been widely utilized. However, polarimetric signals are inevitably distorted by channel imbalance(CI) and interference between multiple channels, resulting in rapidly degraded quality of polarimetric SAR images. Therefore, several polarimetric calibration methods using a polarimetric active radar calibrator(PARC), which can provide different scattering matrices, have been developed. Nevertheless, errors generated by the performance of imperfect PARCs are inevitable, leading to significant errors of estimated CIs. In this study, we propose a framework for calibrating polarimetric SAR images, which consists of two stages: 1) coarse estimation of distortion parameters through a conventional method using three PARCs, and 2) fine estimation of CIs using particle swarm optimization and a single PARC. In simulations using polarimetric SAR images, we observed that our proposed method can more accurately calibrate polarimetric SAR images as compared to conventional methods.