Robust and video-frame rate tracking of planar motion by matched filtering on FPGA

We describe the implementation of a vision algorithm based on matched filtering on an FPGA. Detection of a reference image in an input image must be performed in real-time and robustly against occlusion and change of illumination. Vision algorithm based on matched filtering can perform the detection robustly but requires much computation time since it includes 2D FFT/IFFT and polar transform. In this article, we implement the algorithm on an FPGA so that the detection is performed in real-time and robustly.

[1]  Michel Defrise,et al.  Symmetric Phase-Only Matched Filtering of Fourier-Mellin Transforms for Image Registration and Recognition , 1994, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  D. Casasent,et al.  New optical transforms for pattern recognition , 1977, Proceedings of the IEEE.

[3]  Dana H. Ballard,et al.  Generalizing the Hough transform to detect arbitrary shapes , 1981, Pattern Recognit..

[4]  Hiroyuki Onishi,et al.  Detection of rotation and parallel translation using Hough and Fourier transforms , 1996, Proceedings of 3rd IEEE International Conference on Image Processing.