Abstract
The recognition of colored objects is very important for robot vision in RoboCup Middle Size League competition. Dynamically changing light conditions can cause lots of difficulties to it. This paper describes a robust object recognition method based on our omni-directional vision system for our RoboCup Middle Size League robot-NuBot. The conditional probability density distributions of the YUV values mapping to each color are verified to be Gaussian and the means and variances are obtained by manual calibration. In the image processing, the classifying seeds are selected based on the means and variances, the object regions are grown by the principle that the colors in an object region should be similar, and then the means and variances are updated to adapt to the changing illumination. The experiment results show that the recognition method can be adaptive to light conditions when the illumination is not changed very suddenly and greatly.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have