摘要
全景视频是更先进的虚拟现实技术,用数个摄像机联合拍摄周围的景物,通过全景缝合将多个摄像机所拍摄的视频合为一体,生成全景视频,真实再现周边的实况全景。提出了用于生成全景视频的高性能计算架构,采用图形处理器完成大部分全景视频生成算法,降低了中央处理器的使用个数。应用上,将全景视频与IPTV技术结合,生成的全景视频通过剪裁、压缩编码、传输,发送给远程观众,使远端观众可以自由地观看其感兴趣的区域,产生一种新的全景直播方式,为数字博物馆参观提供了新途径。
Panoramic video is a more advanced virtual reality technology. It uses muhi-camerascheme to capture 180° or 360°surrounding, and subsequently stitches the multiple videos into a wide-angle, high resolution panoramic video, which recreates real environment of the surroundings of the cameras. It proposes a high- performance architecture to panoramic video generation. It adopts graphic processor to perform the most of the work on the panroammic video generating, which drastically reduced the requirement number of the central processors. On the aspect of application, combining the panoramic video with IPTV technology, deliver the panoramic video to the remote watchers through the processes of clipping, encoding and network sending. Owing to panoramic video, the remote viewers have the ability to change their field of vision freely, and they can look wherever they are interested in. This technology brings a new TV live broadcast, which open a new scene of the Digital Museum in the exhibition way.
出处
《北京联合大学学报》
CAS
2015年第3期5-9,共5页
Journal of Beijing Union University
关键词
全景视频
通用GPU
多摄像机
直播系统
Panoramic video
General-Purpose GPU
Multi-camera
Live broadcast system