Benchmarking YOLOv8 for Optimal Crack Detection in Civil Infrastructure

Ensuring the structural integrity and safety of bridges is crucial for the reliability of transportation networks and public safety. Traditional crack detection methods are increasingly being supplemented or replaced by advanced artificial intelligence (AI) techniques. However, most of the models rely on two-stage target detection algorithms, which pose concerns for real-time applications due to their lower speed. While models such as YOLO (You Only Look Once) have emerged as transformative tools due to their remarkable speed and accuracy. However, the potential of the latest YOLOv8 framework in this domain remains underexplored. This study bridges that gap by rigorously evaluating YOLOv8's performance across five model scales (nano, small, medium, large, and extra-large) using a high-quality Roboflow dataset. A comprehensive hyperparameter optimization was performed, testing six state-of-the-art optimizers-Stochastic Gradient Descent, Adaptive Moment Estimation, Adam with Decoupled Weight Decay, Root Mean Square Propagation, Rectified Adam, and Nesterov-accelerated Adam. Results revealed that YOLOv8, optimized with Stochastic Gradient Descent, delivered exceptional accuracy and speed, setting a new benchmark for real-time crack detection. Beyond its immediate application, this research positions YOLOv8 as a foundational approach for integrating advanced computer vision techniques into infrastructure monitoring. By enabling more reliable and proactive maintenance of aging bridge networks, this work paves the way for safer, more efficient transportation systems worldwide.

Data-Driven Method for Enhanced Corrosion Assessment of Reinforced Concrete Structures

Corrosion is a major problem affecting the durability of reinforced concrete structures. Corrosion related maintenance and repair of reinforced concrete structures cost multibillion USD per annum globally. It is often triggered by the ingression of carbon dioxide and/or chloride into the pores of concrete. Estimation of these corrosion causing factors using the conventional models results in suboptimal assessment since they are incapable of capturing the complex interaction of parameters. Hygrothermal interaction also plays a role in aggravating the corrosion of reinforcement bar and this is usually counteracted by applying surface-protection systems. These systems have different degree of protection and they may even cause deterioration to the structure unintentionally. The overall objective of this dissertation is to provide a framework that enhances the assessment reliability of the corrosion controlling factors. The framework is realized through the development of data-driven carbonation depth, chloride profile and hygrothermal performance prediction models. The carbonation depth prediction model integrates neural network, decision tree, boosted and bagged ensemble decision trees. The ensemble tree based chloride profile prediction models evaluate the significance of chloride ingress controlling variables from various perspectives. The hygrothermal interaction prediction models are developed using neural networks to evaluate the status of corrosion and other unexpected deteriorations in surface-treated concrete elements. Long-term data for all models were obtained from three different field experiments. The performance comparison of the developed carbonation depth prediction model with the conventional one confirmed the prediction superiority of the data-driven model. The variable ...