Design of Autonomous Bolt Detection System Based on Quadrotor UAV

In modern railway construction, manual on-site detection of traction power foundation interfaces can be unsafe and inefficient, especially in scenarios involving high piers. To address this issue, an intelligent detection approach has been proposed based on modular design principles. This approach integrates computer vision technology and a PID flight control algorithm to enhance safety and efficiency. The proposed solution includes an autonomous bolt detection system designed and implemented using a Matrice 300 RTK (Real-Time Kinematic) quadrotor UAV (Unmanned Aerial Vehicle) equipped with a high-performance onboard computer, Manifold 2, a visible light camera H20, and a depth camera with a visual laser range-finder. This system replaces the traditional detection method, reducing manual involvement and improving detection accuracy. The proposed system has been successfully applied and demonstrated in the field, resulting in significantly higher detection precision.