Cost Efficient Virtual Environment Generation Framework Using Annotated Panoramic Videos
Cost Efficient Virtual Environment Generation Framework Using Annotated Panoramic Videos
复制标题
DOI：
10.1109/access.2019.2937877
复制
发表时间：
2019-01-01
期刊：
影响因子：
3.9
通讯作者：
Ganz, Aura
中科院分区：
文献类型：
--
作者：
Dong, Hao;Ganz, Aura
在本文中，我们介绍了一个具有成本效益的框架，使一个现实的表示的物理环境，使用全景视频。该框架包括以下主要模块：1）视频捕获和拼接，2）使用计算机视觉技术在全景视频中估计相机轨迹，以及3）使用集成到全景视频中的注释生成虚拟体验。为了量化处理时间方面的框架性能，我们在三个不同的大型环境中使用3.5GHz台式计算机进行处理进行了实验。我们表明，所提出的框架处理时间小于一分钟每米的记录路径在物理环境中，这是一个显着减少相比，传统的3D建模方法。虽然节省了大量的成本，但该框架使用户能够仅沿着记录的路径沿着探索虚拟环境。拟议的框架提供了一个机会，以有限的预算开发应用程序（例如基于模拟的培训），并仍然满足非常准确地表示物理室内环境的要求。
In this paper, we introduce a cost-efficient framework that renders a realistic representation of a physical environment using panoramic videos. The framework includes the following main modules: 1) video capturing and stitching, 2) camera trajectory estimation in the panoramic video using computer vision techniques, and 3) generation of virtual experiences using annotations that are integrated into the panoramic video. To quantify the framework performance in terms of processing time, we conducted experiments in three different large environments using a 3.5GHz desktop computer for processing. We show that the proposed framework processing time is less than a minute per meter of the recorded path in the physical environment which is a significant reduction compared to traditional 3D modeling approaches. While the cost savings are substantial, the framework enables the users to explore the virtual environment only along the recorded paths. The proposed framework provides an opportunity to develop applications (e.g. simulation-based training) with limited budgets and still meet the requirement of a very accurate representation of the physical indoor environment.