Infrared video tracking of UAVs: Guided landing in the absence of GPS signals

Abstract

Unmanned Aerial Vehicles (UAVs) use Global Positioning System (GPS) signals to determine their position for automated flight. The GPS signals require an unobstructed view of the sky in order to obtain position information. When inside without a clear view of the sky, such as in a building or mine, other methods are necessary to obtain the relative position of the UAV. For obstacle avoidance a LIDAR/SONAR system is sufficient to ensure automated flight, but for precision landing the LIDAR/SONAR system is insufficient for effectively identifying the location of the landing platform and providing flight control inputs to guide the UAV to the landing platform. This project was developed in order to solve this problem by creating a guidance system utilizing an infrared (IR) camera to track an IR LED and blue LEDs mounted on the UAV from a RaspberryPI 3 Model B+. The RaspberryPI, using OpenCV libraries, can effectively track the position of the LED lights mounted on the UAV, determine rotational and lateral corrections based on this tracking, and, using Dronekit-Python libraries, command the UAV to position itself and land on the platform of the Husky UGV (Unmanned Ground Vehicle).