Project Details

Abstract

Pavement marking, as a transportation asset, is highly related to safety and mobility but has a short service life. Faded pavement marking presents a significant concern for road users, compromising their ability to navigate safely. Additionally, litter (e.g., vehicle debris) on the roadways poses a substantial hazard that can significantly contribute to traffic accidents. To address these issues, regular inspection and maintenance of the pavement are necessary, including repainting the faded markings and cleaning the litter on the roadways, to ensure the pavement is in good, clean, and safe condition. However, traditional inspection methods still heavily rely on manual efforts, which are subjective, labor-intensive, and time-consuming, which are not suitable for large-scale and frequent implementation. The advancements in artificial intelligence (AI), particularly deep learning and computer vision, have provided new solutions to inspect transportation infrastructure. However, there is limited application in assessing the conditions of pavement markings and identifying roadway litter. Also, counting and locating the identified issues are less involved in prior studies. Therefore, this AI-based project aims to develop a lightweight, affordable, and automated approach to inspect pavement marking conditions and pavement cleanliness, facilitating efficient planning maintenance work and ultimately improving the safety of road users.

Project Word Files