一种新的基于SIFT算法的景象匹配制导方法

A Novel Scene Matching Guidance Method Based on SIFT Algorithm

为有效提高景象匹配制导的性能,在SIFT算法的基础上,结合Canny算子边缘特征检测对其进行改进,提出了一种新的Canny-SIFT算法。先利用Canny算子获取图像的边缘特征,再用SIFT算法对边缘特征图像进行匹配。仿真结果表明,Canny-SIFT算法对亮度、旋转和尺度变化不敏感,有效克服了SIFT算法匹配时间较长的弊端,在大幅提升匹配速度的同时可以提取到更多匹配特征点,提高了匹配正确率,改善了整个景象匹配制导系统的性能。

In order to improve the performance of scene matching guidance a novel Canny-SIFT algorithm is proposed based on Scale Invariant Feature Transform(SIFT)algorithm. The edge feature of the image is obtained by Canny operator, and the edge feature image is matched by SIFT algorithm. The Canny operator is used to extract the edge feature of image at first. Then, the edge feature image is utilized to realize matching. The simulation results indicate that the Canny-SIFT algorithm is insensitive to the change of brightness, rotation and scale, which effectively overcomes the long matching time of SIFT algorithm. More matching feature points can be extracted when the matching speed is greatly improved, and thus the matching accuracy is improved and the performance of the whole scene matching guidance system is improved.