SAR Target Orientation Estimation Based on Manifold Learning

Abstract

Target orientation estimation is an important step in synthetic aperture radar (SAR) automatic target recognition (ATR). In this paper, we propose a SAR target orientation estimation method using manifold learning. We analyze the low dimensional manifold of SAR targets with different orientation, and show that orientation angle can be estimated from the two dimensional embedding. Results in MSTAR data set are presented and show the validity of the proposed method.