图像配准是超分辨率重建、图像融合的基础。提出了非迭代的基于梯度相位相关的图像自动配准方法。首先分析了图像梯度和相位相关的理论,证明了提出方法的正确性;然后给出了图像配准的过程,并利用人工平移图像配准对该方法进行了实验验证。实验表明,该方法配准精度高,能有效地实现图像的自动配准。
The technique of image registration is the basis of super-resolution reconstruction and image fusion. Noniterative automatic image registration method based on gradient phase correlation is proposed. Theoretical analysis is given to show the correctness of the approach based on gradient phase correlation; Image registration process is presented and the method is validated by experiment of shifted image registration. Experiments show that images are registered automatically using proposed approach.