An Efficient Ball Detection Algorithm in Humanoid Soccer Robots

Machinevision system is one of the most important software components for humanoid robots. Humanoid robots obtain important information from its own environment by using this system. In fact, fast and accurate robot reactions in different situations dependent on the performance of this system. Since the main purpose of vision system is detection and determination of object position, one of the most important object that must be detected by humanoid soccer robot is the ball. Robots usually deal with dynamic environments with variable features, which they must be adapted. It’s not desirable to use static algorithms for these kinds of robots, and in this case, it reduces the efficiency. In this paper, a method has been presented in which robot neck position considered to improve detecting the ball. Also the obtained results are implemented on IRC humanoid robot.

کلیدواژه ها:

Humanoid Robot ، Machine Vision ، Soccer Robots ، Real time Processing