Novel Motion Sensing Algorithm for Improving SAR Imaging by Parametric Error Modeling

This paper presents a new motion sensing algorithm for airborne synthetic aperture radar (SAR). Image processing in SAR requires highly accurate and precise motion information which is used for matched filter in range and azimuth compression of the range Doppler algorithm (RDA). If an embedded global positioning system and inertial navigation system (EGI) is used for motion sensing, it can sense the motion more accurately than using INS or GPS alone. However, the discontinuity in the position error occurs because GPS updates the measurement periodically, and this discontinuity significantly degrades the SAR image quality. To resolve this problem, a parametric error modeling (PEM) is proposed to mitigate the discontinuity errors. The PEM is a method that estimates polynomial parameters of INS error compared to EGI which can be assumed by a polynomial in short term. Then, when the error estimated by PEM is subtracted from the INS position, the discontinuity error of EGI can be eliminated. Numerical simulation shows that applying PEM perfectly eliminates discontinuity of EGI output; then, point target simulation using RDA shows improvement for impulse response function quality. Experimental result verifies the above result and demonstrates that the image quality is improved in SAR imagery.