“DREAMS” Drone Research and Affordable Mapping Solutions: Integrating DJI Tello with ORB-SLAM3

The expanding world of drone technology has led to many exciting possibilities in various fields,such as surveillance and monitoring. Equipping drones with different sensors, depending on the use case makes drones invaluable in many applications, especially for environment mapping. Many of these sensors, or the drones equipped with them are costly and high maintenance, which opens up for the creation and implementation of cost-effective and efficient mapping solutions, making it available for a much broader audience, also those that do not have big budgets. By focusing on building and implementing a low-cost mapping solution, this dissertation demonstrates how SLAM technology can be integrated with low-price drones while achieving accurate maps across various scenarios and environments. Through extensive research, a solution to this challenge presented itself with the DJI Tello drone and the ORB-SLAM3 package. Despite its modest specifications, the Tello drone’s user-friendly and affordable nature makes it an ideal candidate for such a system, and with its monocular camera, proved to be highly compatible with the ORB-SLAM3 algorithm. The open-source nature of ORB-SLAM enables customization, therefore making it applicable in a variety of scenarios. The option to change the number of features to be extracted and change the thresholds, allows the system to perform well even in low-contrast areas. During this project, the system has been tested in various scenarios and environments and as a result, had its limitations and challenges showcased such as limited flight time or poor performance in strong winds. Regardless of its limitations, the system proves its ability to create accurate maps and shows how much potential such a system holds. The result was a package solution for performing ORB-SLAM on the DJI Tello drone, and the insights gained from research conducted in this thesis show that it is entirely possible to achieve high-quality mapping even with a limited budget. In conclusion, this project introduces the development of a drone-based mapping system capable of creating highly detailed maps with just a drone equipped with a 720p pinhole camera. It lays the groundwork for future research, continually improving and enhancing the field of low-cost drone-based mapping systems.