This thesis deals with a real time robot vision system which can be directly applicable for grasping a moving object.
The system consists of three main parts;
1) Visual measurement section
2) Manipulator control part
3) interfacing part, especially emphasizing on the realization of following items.
1) Real time visual data processing with 0.2 sec sampling time
2) To grasp the moving object
3) Pattern recognition for simple figured object
4) low cost implementation.
5) Direct applicability to industrial environment.
In addition, on-line manipulator control method is also studied. The performance of system is tested and discussed in the reasonable robotic environment with stepping motor driven belt conveyor.