摘要
采用Harris角点检测算法进行图像特征检测,采用Sift算法中的特征描述方法进行图像特征描述,之后将图像特征点划分为多对多匹配对,根据特征描述值的支持强度不同建立精确的一对一匹配关系.该算法有效地避免了图像特征分布均匀时的Sift匹配效率较低的问题.
Using Harris corner detection algorithm for image feature detection,adopting Sift algorithm for image feature description,the method divided the image feature points into many-to-many matching pairs,and established one-to-one matching according to the different sustaining intension of the feature description.The method effectively avoids the problem of low matching efficiency with Sift algorithm on the condition that the image feature is uniformly distributed.
出处
《测试技术学报》
2009年第3期271-274,共4页
Journal of Test and Measurement Technology
