【摘要】 提出了一种分层迭代红外与可见光图象的配准方法,研究目的是为了求得高精度配准多传感器图象的新方法,传统的基于塔式分解的方法的研究已经取得很多成果,这些方法主要优点是降低了直接相关配准法的计算复杂度,但其缺点在于仅限于同一传感器图象之间的配准。给出了一种基于塔式的改进迭代算法,使其可应用于多传感器图象配准。其原理是,先进行高斯塔式分层由粗到细地提取配准参数,逐步迭代到最大尺度得到高精度的配准。通过大量红外和可见光图象的配准实验验证了此方法的有效性和可靠性。更多还原

【Abstract】 A hierarchical iteration scheme for infrared and visible image registration is presented. The goal is to derive a new method for highly accurate registration of multisensor images. The existing registration algorithms based on image pyramid decomposition have been popular,as they have advantage of low computational complexity. However they can only be applied to images from the same sensor. The proposed new iteration approach can register both visible and infrared images. To capture large displacements, coarse-to-fine hierarchical refinement of the registration parameters is performed based on the Gaussian pyramid. Parameters are propagated on each round of iteration from a low level to a higher resolution to produce more accurate parameters. Experimental results for registration of infrared and visible images show that the proposed scheme is effective and reliable.更多还原