Efficient real-time drone mapping with visual SLAM and Python

Abstract This paper aims to develop an efficient real-time mapping workflow using a drone employing visual simultaneous localization and mapping technology, implemented using Python Programming. The control module is enhanced to guide the drone in exploring an unknown environment, enabling it to perform functionality like a LiDAR sensor but relying solely on its onboard camera to detect walls and features, and dynamically construct a map of the surrounding environment. The approach's accuracy and performance of the reconstructed map are evaluated and compared with the reference map generated using the ground truth. This work demonstrates the feasibility of using computer vision as an alternative to traditional sensing devices for real-time exploration of indoor and complex environments.