摘要： 针对尺度不变特征变换(SIFT)特征匹配算法存在计算量大、复杂度高的问题，提出一种基于图像Radon变换的改进SIFT特征匹配算法。改进算法在图像的SIFT特征点采样区域内作d条不同方向的直线，以d条直线上的图像Radon变换作为SIFT特征向量描述符，降低SIFT特征向量的维数，从而提高特征匹配效率。实验结果表明，改进算法具有较高的匹配精度和较少的匹配时间，适用于虚拟场景漫游或目标识别等实时性要求较高的系统。

关键词: 尺度不变特征变换, 特征点提取, 图像匹配, Radon变换

Abstract: Aiming at the problems of large calculating scale and high complexity in Scale Invariant Feature Transform(SIFT) feature matching algorithm, this paper presents an improved SIFT feature matching algorithm based on image Radon transform. It makes d beelines on different directions in image SIFT feature point zone. Image Radon transform integral values on d beelines are adopted as SIFT feature vector descriptors, it reduces the dimensions of SIFT feature vector to improve the efficiency of feature matching. Experimental result proves that the improved algorithm has higher matching accuracy and needs less matching time, it is quite suitable for high real-time demanded system such as virtual space roaming and target identification.

Key words: Scale Invariant Feature Transform(SIFT), feature point extraction, image matching, Radon transform

中图分类号: 

• TP393