An adaptive color segmentation algorithm for sony legged robots

Abstract

This paper presents an adaptive colour segmentation algorithm for Sony legged robots to play a football game. A Self-Organizing Map (SOM) is adopted to measure the current lighting condition and an Artificial Neural Network (ANN) is implemented to produce a suitable General Color Detection (GCD) table. Off-line learning is conducted in color segmentation in order for Sony-legged robots to adapt to lighting changes in the real world. Experimental results show the good performance of the proposed segmentation algorithm.