Design and Utilization of an Auto-Visual-Inspection Composite System for Suspension Cables with Fast Flaw Identification

The traditional manual inspection of bridge cables has drawbacks such as low detection efficiency and compromised safety. This paper presents an automated visual inspection system for suspension cables, which incorporates a dual-engine wheeled robot with four U-shaped wheels and an offline artificial intelligence (AI)-based apparent flaw identification method. With its capability to overcome significant obstacles, the robot enables a safe, robust, and efficient on-site collection of images highlighting cable flaws. The climbing ability of the robot and its AI-based auto-inspection of cable flaws were experimentally evaluated through trials under various conditions. Indoor experimental results demonstrated the robot’s carrying capacity of up to 6.8 kg, as well as its crossing capability over the obstacle of 6.3-mm in height on the cable’s surface. Furthermore, field trials conducted on suspension cables of arch bridges strongly evidence the effectiveness of the proposed robot, and the utilization of YOLOv7 demonstrates the rapid, autonomous, and accurate identification of flaw features.