Image matching method based on PIU measure and genetic algorithms

A method based on partitioned intensity uniformity (PIU) measure and genetic algorithms is proposed to achieve the high efficiency and the applicability to multi-modal images in image matching. Firstly, based on the study of PIU the improved PIU (PIU′) is chosen as the similarity criterion in image matching. Then, a series of meliorations are adopted for improving general genetic algorithms in finding global optima. The method can be used for multi-modal image matching without image feature extracting, thus reducing the time complexity and avoiding the errors from imprecise features. The combination with modified genetic algorithms reduces the space complexity. Experimental results show that the method can achieve multimodal image matching at the high speed.