基于多图像融合的裸眼3D视频技术研究  

Research on the technology of naked eye 3D stereoscopic video based on multiple image fusion

作　　者：安玲玲[1] 于雷[1] An Lingling;Yu Lei(Minnan University of Science and Technology Department of Electrical and Electronic Engineering, Shishi, Fujian 362700, China)

机构地区：[1]闽南理工学院电子与电气工程学院,福建石狮362700

出　　处：《计算机时代》2018年第12期4-6,11,共4页Computer Era

基　　金：闽南理工学院校级科研基金项目"基于多图像融合的裸眼3D立体视频技术研究"(17KJX034)

摘　　要：裸眼3D视频的制作是裸眼3D显示和播放中的最为重要的环节,现有柱状透镜式的裸眼3D视频在制作过程中,必须在同一时刻对多个视点图像进行拍摄,成本较高,并且制作视频流程较复杂。以人眼多视点观察的相关理论为研究基础,提出一种图像匹配算法,由单个视点图像生成多视点图像,并对多视点图像进行融合处理,实现将双目视频转换为可在柱状透镜裸眼3D显示器上播放的3D视频,从而降低3D视频制作的复杂度和成本。测试结果表明,该算法的视频转换效果符合预期指标,具有较好的推广价值。The production of naked eye3D video is the most important part of the naked eye3D display and play.The existing columnar lens type naked eye3D video must be photographed at the same time,and the cost is high and the video process is more complex.Based on the related theory of eye multiple view observation,an image matching algorithm is proposed,which generates multiple view images from a single point of view image,and converts multiple view image into a3D video that can be played on a columnar naked eye3D display,and could reduce the complexity and cost of3D video production.The test results show that the video conversion effect of the algorithm meets the expected target and has good popularization value.

关 键 词：裸眼3D 双目视频 多视点图像融合 图像匹配 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]