虚实融合场景的快速构建已经成为虚拟现实和数字媒体领域的热点研究课题，它能够克服单独虚拟场景中表示所导致的真实感不足，也能降低虚拟场景中的复杂场景构建的困难，同时解决二维视频场景中三维空间感不足、场景目标信息的分析计算不便等问题。本项目主要研究多视频图像的畸变校正与图像增强、光照一致性融合、基于虚拟场景结构语义特征的场景视频图像目标 三维快速重建、基于GPU集群的虚实融合场景快速计算等关键技术，在分布式虚拟场景上构建一个基于摄像头网络的虚实融合场景快速重建原型系统，在数字战场指挥或者应急响应监控等1-2个典型场景进行验证试用，以验证虚实融合真实感环境构建的高效率和场景动态目标识别与跟踪的可靠性。本项目成果有望在军事仿真 、应急管理与安全监控、数字文化和智慧城市等领域得到应用。

The fusion and reconstruction between virtual environment and video scene has become a hot research topic in the field of virtual reality and digital media,it overcomes the shortage of realism and reduces the complexity of reconstruction in the virtual environment, meanwhile it overcomes the lack of spaciousness and inconvenience of information analysis and computing in video scene. This project mainly focus on the integration of multi-view video distortion correction and illumination consistency, the scene object 3D rapid reconstruction based on scene semantic structures,and the video scene super-resolution reconstruction, and to quickly calculate the 3D reconstruction of the video scenes on GPU cluster, and multi-user multi-resolutional information display on demand based on multi-clues. Then through distributed virtual scene, we build a camera-network based virtual-video fusion system, and apply in 1-2 typical applications in the didital battlefield command and video surveillance of emergency response to verify the features of reality and real-time changes in virtual-real fusion environment. The results will be applied in military simulation, emergency management and security surveillance,digital culture and smart city.