CPS: Synergy: Autonomous Vision-based Construction Progress Monitoring and Activity Analysis for Building and Infrastructure Projects

该信息物理系统（CPS）奖支持研究，使建筑和基础设施建设项目的自动化监控成为可能。施工监控的目的是为开发商、承包商、分包商和商人提供他们需要的信息，以便轻松、快速地做出项目控制决策。这些决策对建设项目的整体效率有直接的影响。鉴于建筑业是一个价值8000亿美元的行业，提高效率可以节省大量成本，对美国经济和社会都有好处。特别是，通过评估完工进度（进度监测）和如何利用建筑资源（活动监测）的自动化工具，可以大大减少建筑成本和交付时间。这些工具将由计算机视觉、机器人技术和施工管理等学科的进步提供。这个项目的跨学科性质将在这些学科之间产生协同作用，并将对工程教育产生积极影响。与行业的合作伙伴关系也将确保这些进步对建筑实践产生积极影响。施工监测的过程包括数据收集、分析和报告。研究将解决现有的科学挑战，使这三种活动自动化。通过空中机器人和摄像头网络录制视频，数据收集将实现自动化。关键的研究目标是推导出保证建筑工地完全覆盖的规划算法，并推导出基于视觉的控制算法，使摄像机能够稳健地放置和检索。分析将通过数字建筑信息模型自动化，建筑资源可以被跟踪。主要研究目标是提高基于图像的重建的效率和可靠性，识别材料属性和几何形状，建立表示建筑活动的形式语言，并扩展基于部件的自动化活动识别方法。报告将通过无处不在的数字建筑信息模型显示实现自动化。研究的主要目标是形式化约束施工本体及其相关的分类机制，并允许对施工进度进行系统的挣值分析。实验验证将侧重于监测建筑物和基础设施系统中下层结构和上层结构骨架元素的施工，以及相关的土方移动、混凝土放置和钢铁安装活动，这些都是建筑项目中常见的。

This Cyber-Physical Systems (CPS) award supports research to enable the automated monitoring of building and infrastructure construction projects. The purpose of construction monitoring is to provide developers, contractors, subcontractors, and tradesmen with the information they need to easily and quickly make project control decisions. These decisions have a direct impact on the overall efficiency of a construction project. Given that construction is a $800 billion industry, gains in efficiency could lead to enormous cost savings, benefiting both the U.S. economy and society. In particular, both construction cost and delivery time could be significantly reduced by automated tools to assess progress towards completion (progress monitoring) and how construction resources are being utilized (activity monitoring). These tools will be provided by advances in the disciplines of computer vision, robotics, and construction management. The interdisciplinary nature of this project will create synergy among these disciplines and will positively influence engineering education. Partnerships with industry will also ensure that these advances have a positive impact on construction practice.The process of construction monitoring involves data collection, analysis, and reporting. Research will address the existing scientific challenges to automating these three activities. Data collection will be automated by recording video with aerial robots and a network of cameras. Key research objectives are to derive planning algorithms that guarantee complete coverage of a construction site and to derive vision-based control algorithms that enable robust placement and retrieval of cameras. Analysis will be automated with a digital building information model with respect to which construction resources can be tracked. Key research objectives are to improve the efficiency and reliability of image-based reconstruction, to recognize material properties as well as geometry, to establish a formal language for representing construction activities, and to extend a parts-based approach for automated activity recognition. Reporting will be automated with a ubiquitous display of the digital building information model. Key research objectives are to formalize a constraint construction ontology with associated classification mechanisms and allow for systematic earned value analysis of construction progress. Experimental validation will focus on monitoring construction of substructure and superstructure skeletal elements in buildings and infrastructure systems as well as the associated earth-moving, concrete placement, and steel erection activities that are common in construction projects.