摘要
图像配准是图像处理的一个重要技术,被广泛的用于遥感图像、医学影像、视频监控等诸多领域中。提出了一种新的配准算法,首先分析两幅图像的联合直方图点集的分布情况,基于此定义了直方图点集的散度公式,并将其作为两幅图片的相似性测量。为加速参数的搜索过程,配准是在小波域内进行的,并使用Powell算法来优化搜索策略。实验证明相对于基于互信息的图像配准算法,本算法参数优化方法选择可以更灵活,时间消耗更少,噪声鲁棒性更优。
Image registration is a kind of important technology of image processing, and is used in the field of remote sensing, medical image, video surveillance, and so on. An image registration algorithm was proposed： the point set distribution of joint histogram was analyzed. Its divergence formula was defined and used as the similarity metric. To speed up searching the registration parameters, all were done in the wavelet field and Powell method was used to optimize this process. Experiments show that this algorithm can apply wider optimization methods and saves more time and has better anti-noise performance compared with the algorithm based on mutual information.
出处
《微计算机信息》
北大核心
2007年第27期265-267,共3页
Control & Automation
基金
国防基础科研项目基金(编号不公开)
关键词
图像配准
直方图散度
互信息
Image registration, Histogram divergence, Mutual information
