摘要
针对灰度和图像特征存在较大差异的可见光和红外图像配准,给出了一种基于归一化互信息与小波变换相结合的快速图像配准算法。实验表明,该算法加快了配准过程,较利用原图像配准速度提高了约一个数量级,且配准结果具有较高的准确性和稳定性。可见该算法适用于多光谱图像配准。
A new algorithm of infrared and visible light image registration based on normalized mutual information（NMI） and wavelet transform is given.Experimental results show that the match time is faster than traditional algorithms′,and the registration results is more accurate and more stable.It is obvious that the algorithm is useful for registration of the infrared and visible light image.
出处
《红外与激光工程》
EI
CSCD
北大核心
2008年第1期164-168,共5页
Infrared and Laser Engineering
关键词
图像配准
互信息
小波变换
Image registration
Mutual information
Wavelet transform
