Robust color classification for global soccer vision

Abstract

The task of global vision module is to extract meaningful data for the strategy decision module. According to these data, the decision-making module estimates the field condition, and then plans strategies to offense or defense. The data extracted should be reliable and accurate for strategy decision module so as to plan efficient tactics. Robust color classification plays a dramatic role in analyzing the scene based on pre-defined color classes. In addition, appropriate color classification can reduce the computational time and improve the reliability of extracted data by eliminating the uninterested background information. In this paper, principal component analysis (PCA) is adopted to seek for a color subspace. In this color space, a color classification model can be constructed straightforward. By using this model, colors slightly varied can be robustly classified.