基于改进的SURF图像匹配算法研究
首发时间:2012-11-19
摘要:针对图像特征匹配算法数据量大、匹配时间长的问题,研究了SURF特征匹配算法,并对其进行了改进。首先在图像的尺度空间中提取SURF特征点,并生成扩展的特征描述向量,然后建立KD-Tree特征结构,采用BBF查询机制进行最近邻查询实现特征点快速匹配。实验结果表明,SURF算法进行特征检测的时间是SIFT算法的1/3;使用BBF进行特征匹配,匹配速度提高了2-3倍。
关键词: 图像匹配 Hessian矩阵 KD树 BBF 最近邻搜索
For information in English, please click here
Image Matching Algorithm Research Based on Improved SURF
Abstract:With the problem of the large data amount and long search time in detecting and matching features of image matching, a new image matching algorithm based on SURF features is presented. Firstly extract the feature points from the scale space of the image,and create extended feature vectors,then build kd-Tree feature structure, match the interest points by using BBF searching method.Experimental result express that, the time of detecting features by SURF algorithm is one third of SIFT algotithm; The matching speed by using BBF searching method is increased 2-3 times.
Keywords: Image matching Hessian matrix KD-Tree Best-bin-first Nearest neighbor searching
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
基于改进的SURF图像匹配算法研究
评论
全部评论0/1000