Riccati observer design for homography decomposition.

This paper presents a novel approach for tackling the classical problem of homography decomposition. The novelty of this paper lies in the design of a deterministic Riccati observer for addressing the homography decomposition problem instead of solving it on a frame-by-frame basis like traditional algebraic approaches. By exploiting the temporal correlation of image sequences and the natural low-pass response of the observer, we believe that the resulting estimates are robust and less prone to measurement noise. Moreover, an extensive observability analysis that points out a sufficient uniform observability condition under which local exponential stability is granted has been carried out. The large domain of convergence and good performance of the proposed observer have been demonstrated through simulation results.