摘要
通过检测物体的特征点来完成两幅图像的配准。根据USAN区域的描述,采用简化了的SUSAN算法检测刚体的特征点,然后通过找到的特征点和刚体形状不变的特性采用特征点之间的几何关系来配准。实验的结果表明,在无旋转条件下的刚体图像配准中,该算法在特征点提取和图像配准过程中速度比较快。
To solve image registration, modified SUSAN algorithm was used to detect the comer points. On the ground of the unchangeable quality of rigid body and the geometry relation between the feature points, the image registration was accomplished. The experiment proves that it is at high speed in the course of feature extraction and image registration.
出处
《计算机应用》
CSCD
北大核心
2006年第10期2380-2382,共3页
journal of Computer Applications