GTD-based 2D state-space method and its application to recover shape from range signals

This paper presents a GTD-based 2D state-space algorithm for recovering shape and motion from range based signals. The GTD-based 2D state-space algorithm estimates the range and the range rate simultaneously in one operation using eigen decomposition, thus effectively solving the aliasing problem of the range tracks when the ranges of different scattering centers are overlapped. Compared with Fourier algorithm, the new method is not limited to identifying scatters at range gates, so it could estimate range and rage rate exactly. Meanwhile, three modified invariant-based shape and motion reconstruction algorithms for range data taken form unknown viewpoints are presented in this paper, including a novel sequential factorization method. Experiments using synthetic data illustrate that the state-space could effectively satisfy the demand of reconstruction algorithms for exact range data.