【作者】 刘鹏；

【导师】 周军； 谢维佑；

【作者基本信息】 上海交通大学 ， 电子与通信工程， 2009， 硕士

【摘要】 红外与可见光传感器是图像系统中常用的两种传感器,对这两种传感器图像进行有效的融合,能够得到更加丰富的图像信息,并具有更高的可靠性,有利于提高对图像信息的分析和识别能力。快速准确地实现图像配准是图像融合的前提。为此,本文提出了一种基于互信息及蚁群算法实现红外与可见光图像配准的方法。论文从对多模态图像配准的认识入门,介绍了基于互信息的图像配准方法,分析总结了多种配准算法的优缺点。研究了基本蚁群算法的原理和工作流程,及其发展现状;并针对红外与可见光的二维图像设计了一种基于最大互信息与蚁群算法相结合的配准算法。采用互信息配准模型,以图像的灰度统计信息为配准依据,用蚁群算法搜索图像间的最优变换参数,并用最大互信息作为目标函数指导最优变换参数的搜索。又针对现有蚁群算法收敛速度慢的缺点,提出了用遗传算法与之动态融合的方法。最后通过实验验证了算法的可行性和稳定性。本论文为全局优化算法在多模态图像配准中的应用进行了一次有益的探索。更多还原

【Abstract】 IR sensor and visual sensor are widely used in image system. The information and robustness of the image can be enhanced with effective fusion of images from the two kinds of sensors. And it can increase the ability of analysis and recognition for image information. Fast and accurate image registration is the key premise of image fusion. An algorithm of IR and visual image registration based on mutual information and ant colony algorithm is presented.This paper starts from the knowledge of the Multi-modality image registration, introduce the method of image registration, and researches the basic principles of the ant group algorithm and work-flow, studies the developing situation of the ant group algorithm. And it also analyzes and summarizes the advantages and disadvantages of some usual registration and computing methods. The paper designs a registration algorithm which based on the maximal mutual imformation and ant colony algorithm method for two 2D IR and visual images. It takes mutual information model, and bases on the intensities of the images, and uses ant colony algorithm to search the best transformation parameters. It also uses the maximal mutual information as the aim function to guide searching the best transformation parameters. And it designs a registration algorithm which links dynamic combination of Genetic Algorithm and Ant Colony Algorithm, for the slow convergence rate of current Ant Colony Algorithm. Finally, the effectiveness of the algorithm is proved by experimentation.The paper also has a useful try for applying the global optimizing algorithms in the Multi-modality images registration.更多还原