Robust isotropic scaling ICP algorithm with bidirectional distance and bounded rotation angle

This paper proposes a new registration algorithm using bidirectional distance with bounded rotation angle to deal with 2D isotropic scaling point set registration problem. In the paper, a new isotropic registration problem based on the bidirectional distance is presented, which is described as a least square registration model with bounded rotation angle. To solve this problem, the proposed algorithm builds up the correspondence and computes the transformation repeatedly, which is similar to the iterative closest point (ICP) algorithm. A closed-form solution of the transformation is obtained in the given interval at each iterative step. The proposed algorithm monotonically converges to a local minimum from any given initial value. Therefore, to obtain the desired results, the boundary of rotation angle and initial value are estimated by the principle component analysis. A series of experiments are conducted to demonstrate that the proposed method is much more robust when compared with the state-of-the-art point set registration methods.