基于视频模型的虚拟现实视频融合系统

The aim of virtual reality integration is to reduce the visual difference of virtual object and real video image. However, it still has several problems, including heavy imagery distortion when the user deviates from the source viewport, and huge modeling cost of virtual scene. For these issues, this paper proposes a novel virtual-reality video fusion system based on video model. We firstly introduce how we register and build video model with interactive operations, then present the fusion method of video and virtual scenes. On the design and implementation stage, we introduce the architecture of our system. Finally, we test our system’s render performance and compare it with two state-of-the-art virtual reality integration methods, which demonstrates that our fusion method is better than other two methods.