改进的SIFT算法  

Improved SIFT Algorithm

作　　者：殷妍[1] 

机构地区：[1]天津大学仁爱学院计算机科学与技术系,中国天津301636

出　　处：《科技信息》2012年第7期185-185,218,共2页Science & Technology Information

摘　　要：目前图像配准的方法中,SIFT特征点匹配算法是目前图像特征匹配研究领域的一个热点问题。其匹配能力较强,可以处理图像间发生平移、旋转、放射变换的匹配,对任意角度拍摄的图像也具备较稳定的特征匹配能力并且极少数的特征点有偏差。因此通过参数的变化对SIFT算法的原理进行了研究以及改进,并用运算时间等数据说明该算法的有效性。Among recent methods of image registration, SIFT features matching algorithm is a hot issue in the field. It has a strong ability to match that it can be dealt with the match when the images are in translation, rotation, radiation ＂transform, for any angle shot images it still has a stable feature matching capacity and it can seldom mismatch. Therefore the principle of SIFT algorithms has been studied and improved by changing the parameters. The data such as the running times is used to prove its efficiency.

关 键 词：图像配准 SIFT算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]