Artificial intelligence to detect bulging of earth retaining structures in transport systems

Abnormal bulging in earth-retaining structures (ERS) of road and rail transport systems potentially poses safety risks and requires a timely, accurate assessment to prevent structural failures. Traditional inspection methods are often labor-intensive, subjective, and difficult to scale. This study investigates the feasibility of using artificial intelligence (AI) techniques to automate the detection of bulging in ERS, an area of research that remains relatively underexplored. The latest version of the object detection technique YOLOv11s has been improved to identify bulging in visual images captured using a vehicle-mounted data acquisition system. Preliminary results demonstrate high accuracy in detecting anomalous bulging in ERS. The findings highlight the potential use of AI-assisted inspection systems to support efficient, scalable, and consistent structural health monitoring of ERS.