摘要
针对目前不同尺寸的多模态图像自动配准方法存在速度较慢的问题,提出一种改进的多模态图像的自动配准方法。对两幅不同尺寸的多模态图像进行小波变换,以分解后得到的概貌图像为待配准图像,以对齐度为适应度函数,利用遗传算法进行迭代搜索,寻找两幅多模态图像的最佳配准位置。实验结果表明,该方法能实现不同尺寸的多模态图像的自动配准,速度较快,准确性高,鲁棒性强。
For the speed of auto registration method for multi-modal images of different sizes is slow, an auto-registration method based on alignment metric is improved. After doing wavelet transformtion for two multi-modal images of different sizes, and taking the facebook image as the registration and the alignment metric as the fitness function, we study the best registration position of two multi-modal images using the genetic algorithm for iterative search. It is proved that auto-registration for multi-modal images of different size can be realized, and has high accuracy, high speed and strong robustness.
出处
《济南大学学报（自然科学版）》
CAS
北大核心
2012年第4期374-378,共5页
Journal of University of Jinan(Science and Technology)
基金
国家自然科学基金(60873089)
山东省教育科学规划课题重点项目(2008ZK0007)
济南大学科研计划(XKY0926)
关键词
多模态图像配准
互信息
对齐度
小波变换
遗传算法
multi-modal image registration
mutual information
alignment metric
wavelet transformtion
genetic algorithm
