静止多目视频拼接技术研究
发布时间：2019-03-19 21:41
【摘要】：随着计算机及摄像机等设备性能的提高,视频获取与处理的成本日益降低,视频拼接渐渐在图像拼接的基础上发展起来。在监控系统中利用视频拼接技术获得一个大场景、高分辨率的视频,便于对所拍摄场景有一个全面的观察,监视。视频拼接在生活中有广泛重要的应用,比如视频会议,虚拟现实等。本文针对多个静止摄像头拍摄的同一场景的视频拼接展开研究,首先介绍了图像与视频的不同,并对图片拼接的流程做了详细的介绍,还重点论述了SURF特征点的提取过程,SURF特征点粗匹配的方法,使用亮度对比符加速粗匹配的方法,采用基于相似度的改进RANSAC算法计算转换矩阵,图像融合,计算融合区域应用到下次的融合中等。本文的创新点在于提出了一种基于相似度的改进RANSAC算法,提出了一种计算图像融合区域的新方法,并提出了一套完整的视频融合方案,通过该方案可以实时的完成多目静止视频拼接,并保证拼接结果的正确性。
[Abstract]:With the improvement of the performance of computer and camera, the cost of video acquisition and processing is decreasing day by day, and video splicing is gradually developing on the basis of image splicing. In the monitoring system, video splicing technology is used to obtain a large scene, high resolution video, and it is convenient to have a comprehensive observation and monitoring of the photographed scene. Video splicing has a wide range of important applications in life, such as video conferencing, virtual reality and so on. In this paper, the video splicing of the same scene photographed by multiple still cameras is studied. Firstly, the difference between the image and the video is introduced, and the flow of image splicing is introduced in detail. The extraction process of SURF feature points is also discussed in detail. The coarse matching method of SURF feature points uses luminance contrast characters to accelerate rough matching. The improved RANSAC algorithm based on similarity is used to calculate the transformation matrix and image fusion. The fusion area is applied to the next fusion. The innovation of this paper is to propose an improved RANSAC algorithm based on similarity, a new method to calculate image fusion region, and a complete video fusion scheme. Through this scheme, the multi-object still video mosaic can be completed in real time, and the correctness of the splicing result can be guaranteed.
【学位授予单位】：北京理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41
本文编号：2443928
[Abstract]:With the improvement of the performance of computer and camera, the cost of video acquisition and processing is decreasing day by day, and video splicing is gradually developing on the basis of image splicing. In the monitoring system, video splicing technology is used to obtain a large scene, high resolution video, and it is convenient to have a comprehensive observation and monitoring of the photographed scene. Video splicing has a wide range of important applications in life, such as video conferencing, virtual reality and so on. In this paper, the video splicing of the same scene photographed by multiple still cameras is studied. Firstly, the difference between the image and the video is introduced, and the flow of image splicing is introduced in detail. The extraction process of SURF feature points is also discussed in detail. The coarse matching method of SURF feature points uses luminance contrast characters to accelerate rough matching. The improved RANSAC algorithm based on similarity is used to calculate the transformation matrix and image fusion. The fusion area is applied to the next fusion. The innovation of this paper is to propose an improved RANSAC algorithm based on similarity, a new method to calculate image fusion region, and a complete video fusion scheme. Through this scheme, the multi-object still video mosaic can be completed in real time, and the correctness of the splicing result can be guaranteed.
【学位授予单位】：北京理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41
【参考文献】
相关期刊论文 前6条
1 张志纯;况立群;韩燮;杨晓文;;一种改进的RANSAC图像拼接算法[J];计算机测量与控制;2014年06期
2 付倩文;何明一;徐晨宇;;一种带预处理的RANSAC图像拼接算法[J];电子设计工程;2013年15期
3 朱永强;汤雄;江雪;;基于MPEG标准的视频压缩技术分析与研究[J];计算机与现代化;2012年10期
4 刘斌;王涛;;一种高效的平面点集凸包递归算法[J];自动化学报;2012年08期
5 谭正华;王李管;陈建宏;黄俊歆;毕林;;基于KD树的露天矿采场示坡线自动生成方法[J];计算机应用;2010年07期
6 金永霞;全新的多媒体信息压缩技术DivX[J];河海大学常州分校学报;2001年02期
相关博士学位论文 前2条
1 邵向鑫;数字图像拼接核心算法研究[D];吉林大学;2010年
2 宋智礼;图像配准技术及其应用的研究[D];复旦大学;2010年
，本文编号：2443928
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2443928.html
