AN ADAPTIVE COLOR SEGMENTATION ALGORITHM FOR SONY LEGGED ROBOTS

摘要

This paper presents an adaptive colour segmentationrnalgorithm for Sony legged robots to play a football game.rnA Self-Organizing Map (SOM) is adopted to measure therncurrent lighting condition and an Artificial NeuralrnNetwork (ANN) is implemented to produce a suitablernGeneral Color Detection (GCD) table. Off-line learning isrnconducted in color segmentation in order for Sony-leggedrnrobots to adapt to lighting changes in the real world.rnExperimental results show the good performance of thernproposed segmentation algorithm.