首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于改进SIFT算法的视频图像序列自动拼接
引用本文:卢斌,宋夫华.基于改进SIFT算法的视频图像序列自动拼接[J].测绘科学,2013,38(1):23-25.
作者姓名:卢斌  宋夫华
作者单位:中国计量学院机电工程学院,杭州,310018
摘    要:本文提出了一种自适应帧采样和限定特征提取区域的拼接方法,根据帧间重叠率和帧间隔建立线性模型,并把各帧图像对准到其前后的关键帧上。在特征点提取方面,提出了一种改进的SIFT算法进行特征点提取,并采用随机采用一致性(RANSAC)方法来更新匹配点,在图像融合中采用线性加权渐入渐出的自然融合算法。实验结果表明:该方法对一般场景能稳定的抽取到关键帧,并进行拼接,取得了较好的拼接效果。

关 键 词:视频图像  重叠率  SIFT  图像融合

Automatic video mosaic imaging based on improved SIFT algorithm
LU Bin,SONG Fu-hua.Automatic video mosaic imaging based on improved SIFT algorithm[J].Science of Surveying and Mapping,2013,38(1):23-25.
Authors:LU Bin  SONG Fu-hua
Institution:(College of Electromechanical Engineering,China Jiliang University,Hangzhou 310018,China)
Abstract:As the existing methods for video image mosaic take high computational costs and huge computation,this paper proposed a method for making panoramas by getting an adaptive frame and matching method of limiting behavior region.According to inter-frame rate and frame overlap interval of linear model,all frames were matched to their latest neighbor key frames.In the process of mosaic,at the aspect of feature extracting,this paper used the improved SIFT feature operator,and the improved RANSAC algorithm to reduce image registration error.This article used the linear weighting fusion algorithm to achieve the desired gradually-enter gradually-leave effect in the overlap image region.Experiments on video sequences showed that this algorithm could extract mosaic key frames with satisfying quality.
Keywords:video image  overlap ratio  SIFT  image fusion
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号