基于改进SIFT特征和图转换匹配的图像匹配算法

针对SIFT特征在纹理丰富的图像中提取较多的伪点和不稳定的点而影响图像匹配的问题, 提出了一种基于Harris阈值准则的局部不变特征图像匹配算法。该算法在提取SIFT不变特征的基础上, 利用Harris阈值准则对所提取到的不变特征进行选择, 剔除了图像区域中大量可区分性较差的特征点, 从而得到了相对稳定和可区分性较好的特征点。其次, 结合不变特征矢量与图转换匹配(GTM)的方法对提取到的稳定特征点进行了精确匹配。实验对比结果表明, 用取得稳定的特征点, 进而结合一种好的匹配策略, 能够更加增强图像匹配的高效性和鲁棒性。

Algorithm of image matching based on improved SIFT feature and graph transformation matching

As the SIFT operator might extract more false keypoints in the image with various texture, which would affect the result of image matching, this paper proposed a new algorithm of image matching based on SIFT local invariant feature of Harris threshold criterion. On the basis of extracting SIFT invariant features, the extracted invariant feature was selected based on Harris threshold criterion. Therefore, there leaved some more robust and well separable features because the worse separable features were rejected in some region of close-grained image. Finally, it used the vector of invariant feature and graph transformation matching method to match accurately. The experimental results demonstrate that the image matching is high efficiency and robust if it combines stable features with a better matching strategy.