A frequency domain method for registration of range data

Free-form 3-D surfaces registration is a fundamental problem in 3-D imaging, typically approached by extensions or variations of the ICP algorithm. This work presents an alternative method for 3-D motion estimation based on the Fourier transform of the range data. The frequency domain techniques for estimating motion parameters are nonfeature-based methods suitable for unsupervised registration of 3-D views. The proposed method can give unsupervised registration of range data within 1° of accuracy of the angular parameters, useful “per se” in applications where this kind of precision is adequate or which can serve as good starting point for the ICP algorithm when a higher precision is needed