基于直线特征匹配的序列图像自动配准

Automatic Registration of Images Sequence Based on Line Matching Approach

  • 摘要: 提出了一种运用相应直线段特征作为控制基础的序列图像自动配准方法。根据相应直线段“共线”原理建立了图像配准模型,该模型的优点是不必精确定位相应直线段的端点,突破了相应特征必须严格“同名”的限制。同时,根据图像配准对控制直线段的要求,分别设计了序列图像中直线段特征的自动提取和自动匹配算法,从而实现了概略对准条件下的序列图像的全自动配准。

     

    Abstract: Traditionally,image registration approaches are classfied to two categories: area-based methods and feature-based methods,where the latter use corresponding features as control elements.An automatic images sequence registration method is proposed which using corresponding line segment features as control elements.According to the collinearity principle of corresponding line segment features,the image registration model is established.With this registration model,people need only to specify corresponding line segments,but does not need to locate their end points.This method break out the limitation about the corresponding features must be the same.Meanwhile,to realize the automatic registration of image sequence,the algorithms for automatic line extraction and automatic line matching is designed,which satisfy the demands of image registration,under the supposition that the images have overlaid roughly.

     

/

返回文章
返回