Optical and SAR Image Registration Based on Multi-Scale Orientated Map of Phase Congruency

Optical and Synthetic Aperture Radar (SAR) images are highly complementary, their registrations a fundamental task for other remote sensing applications. Traditional feature-matching algorithms fail to solve the significant nonlinear radiation difference (NRD) caused by different sensors. To address this problem, robust registration algorithm with multi-scale orientated map of phase congruency (MSPCO) is proposed. First, diffusion scale space established obtain invariance feature points. Compared linear Gaussian space, can better preserve edge texture information. Second, ensure quantity repeatability features, corner points detected on moment congruency, respectively, which foundation next matching. Third, MSPCO descriptor constructed via orientation (PCO). PCO NRD, scales PCOs enhance robustness descriptor. Finally, strategy based an effective ratio proposed, reduces number comparisons among features improves computational efficiency. The experimental results show that proposed method than existing feature-based methods in terms correct matches accuracy. accuracy only inferior most advanced template matching method, within 0.3 pixels, fully demonstrates optical SAR image registration.