Computer vision-based 3D reconstruction and identification of structural defects

As urban development grows, the need for efficient concrete damage inspection has increased, driving interest in automated damage assessment. This research aims to optimise 3D damage acquisition and analysis for concrete structures. It focuses on three objectives: (1) developing cost-effective, high-precision 3D reconstruction methods by reducing errors caused by concrete's colour and shape; (2) improving reconstruction accuracy and flexibility through handheld, structured light-enhanced stereo vision; and (3) designing features that reduce computational demands while ensuring effective damage recognition. The objectives target enhanced accuracy, efficiency, and practicality for real-world damage assessment of concrete structures through ad-hoc optimisations.