An Open Source Vision Pipeline Approach for RoboCup Humanoid Soccer

Abstract

We are proposing an Open Source ROS vision pipeline for the RoboCup Soccer context. It is written in Python and offers sufficient precision while running with an adequate frame rate on the hardware of kid-sized humanoid robots to allow a fluent course of the game. Fully Convolutional Neural Networks (FCNNs) are used to detect balls while conventional methods are applied to detect robots, obstacles, goalposts, the field boundary, and field markings. The system is evaluated using an integrated evaluator and debug framework. Due to the usage of standardized ROS messages, it can be easily integrated into other teams’ code bases.