摘要
为了解决红外焦平面非均匀算法中出现的运动估值问题,提出了一种新的快速图像配准算法。该算法对传统光流方程计算配准精度的方法作了改进,利用基于投影信息的一维光流方程计算小数像素位移。通过与其他算法的配准效果比较,证实该算法具有配准精度高、计算复杂度低等优点。
In order to solve the problem of motion estimation for nonuniformity correction of infrared focal plane arrays, a new fast algorithm for image registration is presented. The algorithm improves the traditional method which calculates the matching accuracy by using of normal optical flow equation and it calculates the fractional pixel shift by using of one dimension projection-based optical flow equation. In contrast to other algorithms for image registration, the algorithm has advantages of high matching accuracy and low computational complexity.
出处
《半导体光电》
EI
CAS
CSCD
北大核心
2008年第4期590-592,596,共4页
Semiconductor Optoelectronics
基金
国家自然科学基金资助项目(60572151)
关键词
红外焦平面
图像配准
小数像素
光流方程
infrared focal plane
image registration
fractional pixel
optical flow equation
