Monitoring scour pits around offshore wind turbine foundations for enhanced maintenance

Monitoring scour pits around offshore wind turbine foundations is crucial for their safety and longevity. Currently popular methods include sonar detection and diver surveys. Sonar offers high-resolution three-dimensional (3D) mapping but is prone to be affected by water turbidity and rough seas especially under extremely turbid conditions. Diver surveys provide detailed inspections but are limited by water depth, weather, and high labour costs. The most crucial point is that neither of them can provide continuous, all-weather monitoring of the scour pits. To address this issue, this paper introduces a new scour pit monitoring technique for integration into a seabed Autonomous Underwater Vehicle (AUV). It includes a novel calibration method for underwater binocular cameras and a precise real-time key point detection method, SimCC-MobileNetV3, for detecting the deepest point of the scour pit. Laboratory tests confirm the technique's accuracy and speed, achieving an impressive overall speed of 9.8 frames per second, with a Root Mean Square Error of merely 0.77 mm.