A robust image tracker based on phase correlation and Fourier-Mallin transform

Abstract

This paper presents an image tracking algorithm for surveillance system, electro-optical tracking system and missile image seeker based on phase correlation and Fourier-Mallin transform. The algorithm consists of an image pre-processing module, a translation estimation module using phase correlation, a fine motion estimation module applied when confidence rate comes from correlation output can not fulfill threshold value, and reference image update module. The fine motion estimation can be used for measuring shift, rotate and scale between a reference and input image based on Fourier-Mellin transform. A sub-pixel registration method is used to enhance estimation accuracy. Proposed algorithm was evaluated its accuracy and robustness using some real indoor and outdoor image sequences.