Competition Analysis System For Soccer Robots Based On Global Vision And Trajectory Restrictions

Hong Liu,Hongbin Zha,Fei Lin
DOI: https://doi.org/10.1109/ICSMC.2004.1401035
2004-01-01
Abstract:A novel system Of automatic competition analysis based on global vision for practical robot soccer games is proposed in this paper to tentatively overcome gaps between simulation competitions and practical ones. Firstly, to improve performance of vision systems and to acquire more reliable competition data, several ideas including color space transformation, sensitivity object processing, local searching combined with global tracking are presented. Secondly, four trajectory restrictions of moving objects are introduced for acquiring more reasonable competition data. Finally, a competition analysis system to support technique statistics, data analysis and interested scenes retrieval is proposed and implemented. Experiments show that the system is compact and effective for quantitative analysis of practical robot soccer competitions.
What problem does this paper attempt to address?