PFI-TT: Using artificial intelligence to improve the accuracy of automated pavement condition data collection

创新技术转化伙伴关系（PFI-TT）项目的更广泛影响和商业潜力在于，通过一种新颖、经济、准确、可靠的自动路面状况数据收集系统，使工程师能够做出合理的维护决策。所提出的技术可以降低成本，提高路面数据收集的质量。新技术的成功应用有望改善运输基础设施的维护，进而改善其他基础设施系统，从而造福社会。该项目的一个社会贡献是使用工业标准获取、获取、管理和公开共享路面破损图像。共享本项目生成的图像数据，最终可以提高路面图像处理能力。此外，该项目将为一名博士后研究员、研究生和本科生提供创业教育和领导力发展机会。拟议的项目旨在建立一个具有成本效益的自动路面状况数据收集系统的原型。为了克服不准确性并降低与现有方法相关的成本，所提出的技术将使用基于机器学习的图像处理算法，该算法比当前基于规则的方法更复杂。此外，拟议的系统预计将通过使用分析能力来识别有问题的数据和减少错误，从而改善数据质量保证。最后，拟议的系统预计将减少昂贵硬件的使用，从而确保路面和基础设施管理的财政可持续性。该系统的核心组成部分是：(1)高性价比的3D（三维）图像采集；(2)基于机器学习的路面图像处理算法；(3)系统的设计和优化驱动的数据质量保证。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 Partnerships for Innovation - Technology Translation (PFI-TT) project is in enabling engineers to make sound maintenance decisions through a novel, cost-effective, accurate and reliable automated pavement condition data collection system. The proposed technology should reduce the costs and improve the quality of pavement data collection. A successful application of the new technology is expected to benefit society by improving the maintenance of transportation infrastructure and, subsequently, other infrastructure systems. One societal contribution of this project is the acquisition, accession, management, and public sharing of pavement distress images using industrial standards. Sharing the image data generated under this project may ultimately improve pavement image processing abilities. Furthermore, this project will provide entrepreneurial education and leadership development opportunities for a postdoctoral researcher and graduate and undergraduate students.The proposed project is aimed at prototyping a cost-effective, automated pavement condition data collection system. To overcome the inaccuracy and reduce the costs associated with existing methods, the proposed technology will use machine learning-based image processing algorithms, which are more sophisticated than current rules-based methods. In addition, the proposed system is expected to improve data quality assurance by using analytical capabilities to identify problematic data and reduce errors. Finally, the proposed system is expected to reduce the use of costly hardware, thereby ensuring the financial sustainability of pavement and infrastructure management. The core components of this system are: (1) cost-effective 3D (three dimensional) image acquisition; (2) machine learning-based pavement image processing algorithms; and (3) systemic design- and optimization-driven data quality assurance.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.