Research on Aircraft Patrol Inspection Method Using UAV Based on YOLO Algorithm

Abstract

At present, the walk - around inspection of aircraft mainly adopts the manual walk-around method. This method is costly and inefficient, and is prone to human factors such as missed inspections and false inspections. Therefore, the research on intelligent appearance surface inspection methods is an urgent task. In this paper, through the collection and processing of aircraft exterior images, the YOLO algorithm is used to learn the key parts of the aircraft's around-the-aircraft inspection, and an aircraft exterior damage image data set is established. This method plans the drone's around-the-aircraft inspection flight route. The drone is used to transmit the real-time images of the aircraft taken during the flight back to the ground station. The ground station uses the YOLO algorithm to identify the damage types and degrees of the received images. This method can replace the method of mounting a high-performance computer on the unmanned aerial vehicle, reducing on-board equipment and improving the reliability of the system. The method proposed in this paper can solve some of the drawbacks of traditional manual visual inspection, and can provide a technical reference for the engineering application of robot-intelligent around-aircraft inspection, which is of great significance for reducing the maintenance and upkeep costs during the aircraft flight-test and operation stages.