I-Corps: Automated Pavement Evaluation Using Advanced Machine Learning

I-Corps项目的更广泛影响/商业潜力在于开发一套软件，使城市和州交通部门（DOTs）等公路所有者机构能够以更严格和更具成本效益的方式管理其路面网络，从而更有战略意义地使用维护和修复资金。该软件套件可以很容易地扩展到包括其他民用基础设施，如机场路面、桥梁、铁路和交通轨道系统。此外，与交通设施相关或在交通设施附近的其他资产也可以被定位、识别和评估。该映射包括道路标志、护栏、油漆标记、公用设施、垃圾箱等，这些都可以作为软件套件的数据可视化平台上的附加层来查看。最后，运输可持续性、弹性和环境影响的措施可以作为可视化软件的另一层进行评估和可视化。这种可视化为业主机构提供了一个公正、直接的平台，帮助他们系统地朝着更具可持续性和弹性的基础设施迈进。I-Corps项目是基于自动路面评估软件套件的开发。该套件将结合几种编码和集成的机器学习和深度学习技术，用于检测和分类关键路面障碍物的范围和严重程度。这些算法旨在提供高度准确、公正的路面状况评估。基于机器学习的模型使用了一个综合的路面图像数据集，该数据集考虑了经过路面专家验证的柔性和刚性路面的20种不同路面损伤。该软件的另一个功能是它能够使用互联网街景图像预筛选项目。这种能力可能允许各机构使用配备了所提供的视频捕捉和道路粗糙度传感器系统的市政车队为自己的道路网络收集数据。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this I-Corps project is the development of a software suite that allows highway owner-agencies such as cities and state Departments of Transportation (DOTs) to manage their pavement networks in a more rigorous and cost-effective manner, enabling more strategic use of maintenance and rehabilitation funds. The software suite can be readily expanded to include other civil infrastructure, such as airfield pavements, bridges, and rail and transit track systems. In addition, other assets associated with or in the vicinity of transportation facilities can be located, identified, and assessed. This mapping includes road signs, guard rails, paint markings, utilities, garbage bins, etc., which can be viewed as additional layers on the data visualization platform of the software suite. Finally, measures of transportation sustainability, resilience, and environmental impact may be assessed and visualized as yet another layer on the visualization software. This visualization provides owner-agencies an unbiased, straightforward platform to assist them in systematically moving towards more sustainable and resilient infrastructure.This I-Corps project is based on the development of an automated pavement evaluation software suite. The suite will incorporate several coded and integrated machine learning and deep learning techniques used for the detection and classification of the extent and severity of critical pavement obstacles. The algorithms seek to provide highly accurate, unbiased pavement condition assessments. The machine learning-based models use a comprehensive pavement image dataset, which considers twenty different pavement distresses for both flexible and rigid pavements as verified by pavement experts. Another capability of the software is its ability to pre-screen projects using internet street view images. This capability may allow agencies to crowd-source data for their own road networks using municipal vehicle fleets equipped with the supplied video capture and road roughness sensor system.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.