一种基于显著边缘的大尺度动车图像匹配算法

Image match plays an important role in the TEDS. Traditional image matching methods are slow with poor accuracy when applied to large-scale EMU image. Based on analyzing the significant edges of the image, this paper proposes an EMU image matching algorithm. Firstly, it calculates the location of the feature region by using the parameters of the TEDS acquisition module. Then, a custom template is used in the preprocessing step to repair the broken edge information. For lines detected by Hough transform, a lines merging algorithm is devised to complete the feature extraction. At last, to enhance the confidence, feature regions will be matched. Through lots of experiments，the algorithm is proved that is can meet the demand of real-time as well as high accuracy. This algorithm is suitable in TEDS.