Fast obstacle detection for UAV using vision (4)
Phone, Thit Htun
Date of Issue2016-12-01
School of Electrical and Electronic Engineering
Automatic obstacle detection and avoidance is a key component for the success of unmanned aerial vehicles (UAVs) in the future. As the payload of UAVs is highly constrained, cameras are attractive sensors because they are both lightweight and provide rich information about the environment. In this paper, the student presents an approach that allows a UAV to detect obstacles with horizontal scan using a stereo camera mounted on the UAV. While the UAV is hovering, the stereo camera captures stereo image sequences in its forward direction. The captured stereo images are sent to ground control station for processing. Semi-Global Block-Matching algorithm is used as a matching method for those input stereo image pairs and it outputs a disparity map. From this disparity map, the depth information or distance to the obstacles can be obtained, hence it allows the UAV to detect obstacles in a specific range in its flight path. In this project, the obstacles in up to 25m radius around the aircraft are detected. This paper describes the theory behind the stereo obstacle detection system, the approach and results, and demonstrates the obstacle detection accuracies of the system.
DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems
Final Year Project (FYP)
Nanyang Technological University