Conventional image matching method is a correlation matching method of ergodic region, and much time should be wasted on non-optimal matching points. It is not only waste matching and calculation time but also lead to get non-matching points. Based on Multiresolution Pagoda Structure Algorithm (MPSA) and the genetic algorithm and projection algorithm, a new image matching method (GMPT) is presented.
The algorithm can realize fast image matching on big template and the higher precision of matching. A contrast experiment isconducted and the result demonstrates that the algorithm can reduce the matching time sharply. So the algorithm has been applied to the image matching system.