一种基于角点特征的图像自动配准方法

A Method of Image Automatic Registration Based on the Characteristics of the Image’s Corners

针对多幅图像之间的图像配准问题,该文提出了一种新的图像自动配准算法。该算法应用Harris角点检测方法获得角点信息。在匹配过程中,采用圆形区域和双向相关系数法进行相似性度量。采用圆形区域,很好的解决了旋转问题;采用双向相关系数法,更加保证了匹配的精确度,减少误匹配率。初步实验结果表明:该方法可以高效短时地实现图像间的自动配准。

For the problem of image registration among multiple images,a new automatic image registration algorithm is described in this paper. The algorithm uses Harris correr detection method obtaining comers information. Using a circular area and the method of bidirectional correlative coefficient measures the similarity of the images in the matching process. The use of a circular area is a wonderful solution to the issue of rotation;the use of the method of bidirectional correlative coefficient, more ensure the accuracy of the match, reduce the rate of false match. Preliminary experimental results show that： the method can implement the automatic registration among images highly effective and short-tenn.