Predictive and Precision Pavement Maintenance Methodology Utilizing Multi-Temporal Pavement Images

Abstract

In recent years, pavement imaging systems, such as three-dimensional (3D) laser technology, have been widely adopted by state highway agencies (SHAs) for network-level pavement condition evaluation. This technology enables the collection of high-resolution pavement surface images at highway speed. Despite some SHAs collecting 3D pavement images for several years, these multi-temporal images have not yet been fully utilized to enhance pavement maintenance strategies at the project level. Therefore, this study aims to develop a predictive and precision pavement maintenance methodology by fully leveraging multi-temporal pavement images to advance cost-effective maintenance, allowing expensive treatments (e.g., deep patching) to be applied to isolated pavement spots at the optimal time. To achieve this, a data preparation pipeline was first developed to measure cracking severity and its deterioration for each pavement section. Then, a two-by-two matrix was created to prioritize pavement sections according to both cracking severity and deterioration rate. Finally, using deep patching as an example, a treatment planning optimization function was developed to strategically arrange treatment locations, considering construction operation constraints (e.g., a minimum deep patching distance of 10 m). A case study conducted on six years of data from a 5.8 mi section of US-80 near Savannah, Georgia, demonstrated the feasibility of the developed predictive and precision maintenance strategy, achieving more cost-effective maintenance.