大量实验表明,提出的方法对同种传感器或者不同传感器图像之间的配准都是有效的。

提出了一种新的基于角点检测的图像配准方法,其核心思想是采用一种快速的

基于图像灰度的角点检测新算法,通过沿圆弧曲线扫描获取角点信息,然后根据这些

角点信息建立图像间角点的对应关系,并由此得到初配准参数,最后通过迭代过程以

提高配准的精度. 理论分析和实验结果表明,该算法对图像间的旋转角度没有限制,

配准精度高而且计算量较小.

-Experiments show that the proposed method on the same kind of sensor or between the different sensor image registration is valid. A new corner detection based image registration method, the core idea is to use a fast gray-based new algorithm for corner detection, scanning through the curve along the arc for corner information, and then based on these corner information to establish image correspondence between the corners, and thus be the beginning of registration parameters, and finally through the iterative process to improve the accuracy of registration. Theoretical analysis and experimental results show that the algorithm is rotation between the image does not limit , registration accuracy and less computation.