A Visual Servoing Quadrotor UAV for Contact-Based Inspection of Wind Turbine Blades

With the increasing demand for safe and efficient wind turbine maintenance, especially for tasks involving physical contact such as grounding resistance measurement, there is an urgent need to replace traditional manual inspection methods with autonomous robotic systems. This paper presents a quadrotor UAV system designed for contact-based inspection of wind turbine blades. To address the challenges of maintaining stable contact with curved and oscillating surfaces, a hybrid method combining the You Only Look Once (YOLO) algorithm with a Kalman filter is proposed, enabling robust feature detection and continuous state estimation under dynamic conditions. Building upon this perception module, an image-based visual servoing (IBVS) strategy is developed to achieve precise motion control based on real-time visual feedback, ensuring reliable and consistent contact between the inspection probe and the blade surface. The proposed system is validated on a ROS and Gazebo co-simulation platform. Simulation results demonstrate that the method achieves high tracking accuracy and real-time responsiveness, making it a promising solution for autonomous contact-based inspection of wind turbine blades.