基于SURF特征的五金件图像匹配算法研究

Research on Image Matching Algorithm of Metal Pieces Based on SURF Feature

针对传统图像匹配算法进行特征检测和匹配搜索计算量大、耗时长等问题,提出了一种基于SURF特征的图像匹配算法,并以五金件为实验对象进行了测试。该算法首先提取获得五金件图像的SURF特征,并形成特征描述向量,再建立KD-Tree索引,最后通过KD-Tree算法完成特征匹配操作。结果表明,该算法不仅解决了五金件在发生光照、旋转、尺度等变化后的识别问题,而且匹配准确性高、匹配速度快、实时性好,具有较高的实用价值。

In order to solve the problems of large amount of computation and long time in traditional image matching algorithm, an image matching algorithm of the metal pieces based on SURF feature is researched and tested. Firstly, the algorithm is used to extract the SURF feature and generate feature vectors, and then the KD-Tree index is set up for feature description vectors. Finally, the feature matching operation is completed by the KD-Tree Nearest Neighbor search. The experimental results show that this algorithm can greatly improve the matching speed of feature vectors and has high matching accuracy, which solves the identification of metal pieces in rotation, illumination and scale change and has great practical value.