Abstract: In order to accomplish the fast and precise image matching，this article presents a method based on edge light characteristic of the image.Firstly use LOG operator to pick up the edge information and use Hough transform to pick up the main lights from edge information.Then define similar measure between the two images by light characteristic，and use this measure to estimate rotated angle and find out corresponding light groups.Finally，pick up matching control lights from these lights，calculate matching parameters，and match the images.This paper verifies the validity of the algorithm through with experiments.

Key words: image matching, edge detection, light detection

摘要： 为实现高效率、高精度的图像配准，提出了一种利用图像边缘直线特征的图像配准算法。首先利用LoG算子提取参考图像和待配准图像的边缘信息，利用Hough变换提取图像边缘信息中的主要直线；然后以直线特征定义了参考图像和待配准图像的相似性度量，以此度量估计旋转角度，找出直线组的对应关系；最后在这些直线组中选取配准控制直线，计算配准参数，对图像配准。经过实验，验证了该算法的有效性。

关键词: 图像配准, 边缘检测, 直线检测