Stereo vision-based Collision Avoidance for Unmanned Systems

In the unmanned systems, the intelligent navigating technologies are very important. For this purpose, we introduce a navigation algorithm that includes collision avoidance using a stereo camera. To accurately avoid the obstacles, we propose the object detection, the navigable space searching and the efficient collision avoidance real-time algorithms. In this article, we focus on an unmanned aerial vehicle (UAV) system. The UAV basically flies along the global path towards the goal point entered by the user. If it is not possible, UAV return by performing the proposed local path planning algorithm. The generated local path is based on the new waypoint, and it safely flies to the final goal point using the calculated flight direction and distance information between waypoint and waypoint.