Terrain Matching by Fusing HOG With Zernike Moments

A new matching algorithm is proposed by fusing the histogram of oriented gradients’ and the Zernike moments’ descriptors extracted from the real-time synthetic aperture radar (SAR) images and the real-time elevation maps (REMs), respectively. Both the SAR images and the REMs are acquired by an on-board interferometric SAR simultaneously. The two reciprocal descriptors are fused by weighting their Canberra distances. The numerical experiments demonstrate the advantage of the proposed matching algorithm preliminarily.