Design of dense, accurate stereo maps for fast maneuvering of unmanned aerial vehicles

Abstract

In recent times, unmanned aerial vehicles (UAVs) are popular for several applications like rescue, surveillance, mapping, and so on. However, slow flight motion of Quadrotor UAVs is still a challenging issue to overcome. Although there exist several algorithms for the motion estimation and path planning of UAVs, most of them cannot be applied for fast flight in cluttered urban and forest environments. Many navigation systems based on laser scan matching have been demonstrated for the use on Quadrotor UAVs. Nevertheless, keeping in mind that the UAV is to fly at high speeds (5–10 m/s), an alternative for a heavy laser scanner would be a light-weight stereo camera. On the other hand, the main disadvantage for using stereo camera is that the depth map generated is often sparse and noisy, which is the bottleneck for obstacle detection and path planning. Therefore, a segmentation-based filter has been designed to overcome this problem without being dependent on different scenes and lighting conditions. The proposed filter has been tested on publicly available stereo images as well as data generated from our UAV cameras.