Accurate Landing System of Electrical Power Plant Inspection UAV Based on Vision Guidance

Due to the complex terrain of the electrical power plant, it is difficult for the inspection UAV to distinguish the landing site from the environment and achieve accurate landing in the process of inspection. Aiming at the above problems, this paper designs an accurate landing system based on vision guidance for the inspection UAV. Firstly, the black circular block is designed as the landing sign. Then, vision sensors and embedded processing units on the UAV are utilized to acquire images of the ground in real time, and the image filtering and the adaptive threshold segmentation are used to separate the landing sign from the background. Next, the position of the landing sign is detected by the ellipse detection algorithm, and the UAV is guided to land according to the position of landing signs. Finally, the autonomous landing experiments under different scenarios are carried out, the experimental results show that the system proposed in this paper has stable and reliable performance. Among them, the maximum landing deviation of the system is 11.5 cm, and the autonomous landing time is stable within one minute, which meets the requirements of specific projects.