M. Jamzad, A.H. Khodabakshi, and V.S. Mirrokni (Iran)
Omnidirectional vision, Localization, Object detection, Vision system, Middle-size RoboCup
In this paper, we describe a design of an omnidirectional viewing system for mobile soccer robots in middle size RoboCup. This system that uses a hyperbolic mirror is used for object detection, localization and also collision avoid ance. In addition, the hyperbolic mirror is used as a global viewing device to construct a world model for the soccer field. This world model contains information about the position and orientation of the robot itself and the position of other objects in a fixed coordinate system. In addition, a color based fast object detection method is introduced. It detects objects by searching in a set of jump points looking for colors that match that of the desired object. Using the angle of several fixed landmarks on the field with respect to the robot, we introduce a fast and robust localization method. Our localization method is based on line detec tion using clustering and Hough transform method. This viewing system is combined with a front view that uses a plain CCD camera. This combination provided a total vi sion system solution with satisfactory results and near real time speed.
Important Links:
Go Back
