计算机视觉中基于多照片的同名点自动匹配

对多张图像上同名点的自动匹配算法进行了分析,提出了利用多张图像上的核线约束来实现同名点的匹配。首先,计算出一个待匹配点在另外两张图像中的核线,在核线附近确定出概略匹配点。然后,再次利用核线约束特性在确定的概略匹配点之间确定出唯一的匹配点。最后,用该方法对多张图像间的同名点进行了匹配试验,实验结果表明,该方法效果良好,具有较好的实用价值。

Matching Homologous Points Based on Photographs in Computer Vision

After researching on methods of automatically matching homologous points among images,the paper put forward a method to match homologous points by using the specialty of epipolar lines in images.Firstly,the epipolar lines of one unmatched point were calculated in other two pictures.And crude matched homologous points were found by the specialty of epipolar lines in images.Then the accurate homologous points on the other two images were found between the crude matched homologous points by making their epipolar lines.Finally an experiment result proved that the method was very effective and has practical value.