Object Detection and Location Estimation using SVS for UAVs

In this paper a distinct system for object detection and measuring distance to an object for an Unmanned Aerial Vehicle (UAV) utilizing stereo vision sensor has been proposed. The depth information which is acquired from stereo vision sensor plays a vital role in collision avoidance applications. The system uses a Hexacopter with on board webcams interfaced to a Raspberry pi board to capture video. In our proposed algorithm, the UAV travels in a predefined path and tries to detect objects at each waypoint. Thereafter it computes distance from the waypoint to an object and goes to the new waypoint (i.e. object location) with the assistance of GPS and IMU data. The image processing algorithms are implemented in MATLAB.