基于控制点的多光谱图像自动配准研究

针对同类传感器多光谱图像的配准,提出了一种快速有效的图像自动配准方法。该方法采用Forstner算子分别从参考图像和配准图像提取控制点,再利用基于灰度相关的算法匹配同名像点,然后利用同名像点求解仿射变换参数,实现多光谱图像的配准。通过液晶光谱相机成像实验对该方法进行验证,结果证明了该方法处理速度快,配准精度高。

Automatic Registration for Multi-spectral Images Based on Control Points Algorithm

To process the multi-spectral images registration of the same sensors,a fast and effective auto matic image registration algorithm is presented.The control points of the reference image and matching image are extracted by Forstner operator,the corresponding image points are matched by the algorithm based on the gray correlation,and then the parameter of the affine transform is calculated according to the corresponding im age points.The multi-spectral images registration is achieved.The registration result of multi-spectral images of the liquid-crystal spectral camera proves that the algorithm is efficient and accurate.