Quantcast

Obstacle detection system based on colour segmentation using monocular vision for an unmanned ground vehicle

Research paper by Auday Al-;Mayyahi, Weiji Wang, Phil Birch, Alaa Hussien

Indexed on: 13 Jul '18Published on: 08 Jul '18Published in: International journal of computational vision and robotics



Abstract

An obstacle detection algorithm is introduced for aiding the navigation of unmanned ground vehicles (UGV). Coloured obstacles are placed randomly in an indoor environment. The coloured obstacles are detected, analysed and processed using a proposed monocular vision algorithm. A camera calibration is conducted to determine the relative position and orientation of the UGV with respect to the obstacles based on intrinsic and extrinsic matrices to form a perspective projection matrix. The field geometry is used to obtain a mapped environment in the world coordinates. Our obstacle detection algorithm is proposed to identify the existence of the obstacles in the field. Using bounding boxes around the detected obstacle allows the determination of the obstacles locations in a pixel coordinate frame. Thus, the depth perception is determined by using the pixel coordinates and the camera projection matrix. Real-;time experiments are carried out to demonstrate the validity and efficiency of the proposed algorithm.