A review on computer vision based defect detection and condition assessment of concrete and asphalt civil infrastructure

Reviews computer-vision defect detection and condition assessment for reinforced-concrete bridges, precast tunnel linings, buried concrete pipes and asphalt pavement. It concludes that most relevant defects can be detected automatically, but automatic extraction of defect properties and condition rating had not been achieved. Open problems are laid out for both civil engineering and computer science readers.

Key findings

• Detection of most relevant defect types was already automatable
• Automatic retrieval of defect properties and condition assessment was not yet achieved

Limitations

• Abstract only reviewed
• Not UAV-specific; predates widespread drone use
• Literature up to about 2014