Silva-2011e
From IEETA
Conference proceedings article
| Title | Aerial ball detection in RoboCup Middle Size League using polar histograms |
|---|---|
| Author | João Silva, Mário Antunes, Nuno Lau, António J. R. Neves, Luís Seabra Lopes |
| Booktitle | Proceedings of RecPad2011 |
| Address | Porto, Portugal |
| Volume | |
| Pages | |
| Month | October |
| Year | 2011 |
| Group | |
| Group (before 2015) | Transverse Activity on Intelligent Robotics, Embedded Systems, Computing and Control Laboratory |
| Indexed by ISI | No |
| Scope | National |
In a soccer game, the ball is the most important element of the game, as it is the instrument used by the players to produce the results of the game. In robotic soccer, the robots need to know where it is to be able to play. When building the representation of the environment, the ball information is one of the main issues, and thus its detection and validation is a priority. This paper briefly presents the visual detection of the ball, independently of its surroundings, using a single camera pointing forward. The objective is to merge this information with the one obtained by the robot main sensor, an omni-directional camera. The ball seen with the frontal camera, contrary to the one seen by the omni-directional one, is not surrounded by a semi-controlled environment (on the ground, we know that the ball is surrounded by a limited set of colors). The approach consists of a hybrid color/shape algorithm for detection of the ball in the camera image. The preliminary results obtained showed a good performance.