Edge preserved extrapolation method for full polarimetrie RPM imaging with UWB radars

摘要

Ultra-wideband (UWB) radar system has a great advantage as short range 3-dimentional(3-D) sensor with high range resolution being suitable, for target recognition in optically blurred situation. As one of the most efficient 3-D imaging approaches, the range point migration(RPM) method has been developed, which specifies target boundary extraction using measured ranges. The RPM method stills suffers from insufficient imaging area along cross-range direction, and this incurs the difficulty for target recognition using the reconstructed 3-D image, which is an essential problem in radar imaging. To tackle with this problem, an ellipsoidal aggregation based image extrapolation method, incorporating full polarimetric datasets and RPM imaging, have been already proposed. While this method retains an accurate extrapolation for smooth surface target, it suffers from inaccuracy for target shape with edge. Thus, this paper introduces an edge preserved algorithm based on eigenvector decomposition for quasi Hessian matrix obtained by RPM image. The results from finite-difference time-domain (FDTD) based simulations demonstrate that our method effectively expands a target image without sacrificing an accuracy around edge area.