摘要
针对图像匹配在不同的光照、表情等背景下存在可靠性及有效性较低的问题,本文提出一种改进的SIFT特征的图像匹配算法。算法通过构建高斯差分尺度空间以保留原始图像信息并增加特征点数量,计算相邻点及上下尺度空间对应点以寻找极值点,调整Harris Corner检测器参数以实现删除低对比度的关键点和不稳定的边缘响应点,利用关键点邻域像素的梯度方向分布特性,为每个关键点指定128维的方向参数,并进行关键点描述子的生成,最后将两幅图片的描述子进行匹配并得到结果。实验结果表明,该算法在不同背景下能有效完成图像匹配的计算,满足应用软件的时效性要求。
In view of the problem of low reliability and effectiveness of image matching under different lighting and expression backgrounds,this paper proposes an improved image matching algorithm with SIFT features.The algorithm constructs a Gaussian difference scale space to retain the original image information and increase the number of feature points,calculates adjacent points and corresponding points in the upper and lower scale spaces to find extreme points,and adjusts the parameters of the Harris Corner detector to remove the key points of low contrast stable edge response points.And it uses the gradient direction distribution characteristics of the pixels in the neighborhood of key points to specify 128-dimensional direction parameters for each key point,and generate key point descriptors.Finally,the descriptors of the two pictures are match and result is obtained.Experimental results show that the algorithm can effectively complete the image matching calculation under different backgrounds and meet the timeliness requirements of application software.
作者
胡柳
邓杰
肖瑶星
卢艳芝
曾蒸
HU Liu;DENG Jie;XIAO Yaoxing;LU Yanzhi;ZENG Zheng(Hunan College of Information,ChangSha 410200,China)
出处
《智能计算机与应用》
2020年第8期231-233,共3页
Intelligent Computer and Applications
基金
湖南省教育科学“十三五”规划2018年度教育考试研究专项课题(XJK018JKB023)
作者简介
胡柳(1988-),男,硕士,讲师,主要研究方向:图形图像处理、网络化软件开发。
