Automated Low-Cost Change Detection of Road Infrastructure Assets Using Remote Sensing and AI

使用遥感和人工智能对道路基础设施资产进行自动化低成本变化检测

• 批准号: 576485-2022
• 负责人: Gargoum, SulimanSA
• 金额: $ 3.61万
• 依托单位: University of British Columbia
• 依托单位国家: 加拿大
• 项目类别: Alliance Grants
• 财政年份: 2022
• 资助国家: 加拿大
• 起止时间: 2022-01-01 至 2023-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=743764
• 关键词: Automated; Low; Cost; Change; Detection

Highway infrastructure across Canada is a key national asset that is critical to the mobility and economic prosperity of Canadians. Unfortunately, maintaining highway infrastructure is a costly and resource draining endeavor. Provincial transportation agencies dedicate billions of dollars to the process on an annual basis. In current practice inspecting the conditions of highway infrastructure assets is a tedious process whereby a crew of at least two individuals drive thousands of kilometers of road at low speeds to identify assets' where the conditions have changed since the initial installation. For primary highways the same highway is driven every day and crew stop by at assets in search for any changes in the condition. To help automate the process and improve its efficiency and accuracy, this project will involve developing a low-cost framework that uses remote sensing technology to create virtual 3D models of infrastructure assets. Algorithms that employ principles of machine learning and Artificial Intelligence will then be developed to process the 3D models and provide inspection crew with information about the changes in asset conditions without the need to visit each asset during long and tedious site visits. The results will be shared with the project partner who will work on large scale deployment of the developed technology to improve highway maintenance practice across Canada. This will cause significant improvements in the way infrastructure asset inspections are conducted resulting in safer and more resilient roads at a significantly lower cost to tax-payers.

加拿大的公路基础设施是一项重要的国家资产，对加拿大人的流动性和经济繁荣至关重要。不幸的是，维护高速公路基础设施是一项成本高昂且耗费资源的奋进。省级交通机构每年为这一进程投入数十亿美元。在目前的实践中，检查公路基础设施资产的状况是一个繁琐的过程，其中至少两个人的工作人员以低速行驶数千公里的道路以识别自初始安装以来状况已经改变的资产。对于主要公路，每天都在同一条公路上行驶，工作人员在资产处停下来寻找条件的任何变化。为了帮助自动化这一过程并提高其效率和准确性，该项目将涉及开发一个低成本的框架，该框架使用遥感技术来创建基础设施资产的虚拟3D模型。然后将开发采用机器学习和人工智能原理的算法来处理3D模型，并为检查人员提供有关资产状况变化的信息，而无需在漫长而繁琐的现场访问中访问每个资产。研究结果将与项目合作伙伴分享，后者将致力于大规模部署所开发的技术，以改善加拿大各地的公路养护实践。这将大大改善基础设施资产检查的方式，从而以更低的成本为纳税人提供更安全，更有弹性的道路。