Abstract: Because of the poor light adaptability of the robot vision system, the algorithm of SOM (Self-organizing Map) was applied to the color classification of the robot vision system, owing to its good capacity of clustering. The vision system  has the ability to learn and to adapt to the changes in the environment. The simulation results show that the vision system with the algorithm of SOM has improved the color recognition rate.


Key words: machine vision; algorithm of SOM; clustering; color classification; color recognition rate