Protocol based pavement cracking measurement with 1 mm 3D pavement surface model

Cracking is one of the most significant pavement distresses. Accurately measuring cracking condition from the field has long been a major task for highway agencies. However, due to technical limitation and outdated protocols, the automated cracking measurement has been inconsistent, unrepeatable, and unreliable during the past decades. Recently, a breakthrough in cracking data collection was made possible through the emerging 3D vision technology. The PaveVision3D Ultra system is capable of acquiring 1 mm 3D pavement surface data at highway speed. Meanwhile, new AASHTO cracking protocols are released to support fully automated cracking characterization. In the light of such developments, this dissertation research focuses on exploring a variety of cracking characteristics with the 1 mm resolution 3D pavement data. The specific tasks performed in this dissertation include (1) a comprehensive review on past and current practices of cracking measurement; (2) an evaluation of AASHTO protocol PP67-10 and PP68-10; (3) development of a framework to accurately identify the lane region from the full frame pavement image; (4) a development of critical methods for implementing PP67-10 protocol based cracking characterization, including cracking blob extraction, measurement of cracking width and length, and cracking classification; (5) a proposal and comparison of two advanced methods for deriving cracking width distribution and discussion of their potential applications; (6) a development of a multi-layer index system for PP67-10 based comprehensive cracking evaluation. The outcomes of this research are promising to improve the state of the art of cracking measurement and can be directly applied by highway agencies for advanced pavement cracking evaluation. Recommendations for future research are proposed.