Real-time volumetric video recording of large capture areas

该项目的目标是建立一个能够实时记录快速移动内容的照片级逼真3D模型的系统，最初的重点是体育赛事。当与虚拟现实(VR)、增强现实(AR)或混合现实(MR)硬件配合使用时，这些记录将为观众提供身临其境的全息图式事件渲染，让他们舒舒服服地呆在家里。这项技术(称为体积捕捉)将使用位于事件周围的多个非侵入性传感器来生成3D重建。虽然这个项目的重点将是体育，但这项技术将适用于广泛的现场活动。目前，录制大面积(拳击场大小或更大)的体视频需要处理时间，无法进行直播。为了使我们能够实时记录和分发体育赛事，我们将开发高度优化的算法和数据结构，以及深度学习驱动的系统，以在大范围内实时创建快速移动的事件的体积视频。我们还将与我们的合作伙伴公司合作，开发便携式、易于安装和非侵入性的摄像设备，以免限制观众的视线。

The aim of this project is to build a system capable of recording photorealistic 3D models of fast moving content in real-time, with an initial focus on sports events. When paired with virtual reality (VR), augmented reality (AR) or mixed reality (MR) hardware, these recordings will provide viewers with immersive, hologram-style renderings of events from the comfort of their own homes. The technology (known as volumetric capture) will use multiple, non intrusive sensors positioned around an event to generate the 3D reconstruction. Although the focus of this project will be on sports, the technology will be applicable to a wide range of live events. Currently, recording volumetric video of large areas (as large as or greater than the size of a boxing ring) requires processing times which prohibit live broadcasting. To enable us to record and distribute sports events in real-time, we will develop highly optimised algorithms and data structures alongside deep learning powered systems to create volumetric video of fast moving events over large areas in real-time. Working with our partner company, we will also develop camera equipment which is portable, easy to set up, and non-intrusive so as not to restrict the view of audience members.