Monocular Vision Aided Autonomous UAV Navigation in Indoor Corridor Environments

Deployment of autonomous Unmanned Aerial Vehicles (UAV) in various sectors such as disaster hit environments, industries, agriculture, etc., not only improves productivity but also reduces human intervention resulting in sustainable benefits. In this regard, we present a model for autonomous navigation and collision avoidance of UAVs in GPS-denied corridor environments. In the first stage, we suggest a fast procedure to estimate the set of parallel lines whose intersection would yield the position of the vanishing point (VP) inside the corridor. A suitable measure is then formulated based on the position of VP on the intersecting lines in reference to any of the image boundary axes. The knowledge of VP location alongside the formulated mechanism govern the necessary set of commands to safely navigate the UAV avoiding any collision with the side walls. Furthermore, the relative Euclidean distance scale expansion of matched scale-invariant keypoints in a pair of frames is taken into account to estimate the depth of a frontal obstacle; usually a wall at the end of the corridor. However, turbulence in the UAV arising due to its rotors or other external factors such as wind may introduce uncertainty in depth estimation. It is rectified with the help of a constant velocity aided Kalman filter model. Necessary set of control commands are then generated to avoid the frontal wall before collision. Exhaustive experiments in different corridors reveal the efficacy of the proposed scheme.