Study on the target positioning for an Omni-directional 3 DOF mobile manipulator based on machine vision

The omni-directional mobile robot with multi DOF, because the operation posture and operation accuracy of the manipulator can be better controlled in a narrow or crowded workplace compared with the general manipulator, is getting more interested in practical applications. The present problem is to improve its flexibility for operating multiple different targets. Target recognition with image processing is an effective solution. Based on the image processing, the position and posture of the target can be determined. Then the signal will be sent to the arm control system. In this paper, the illumination conditions, distortion, etc. are studied in the target recognition. The target position with image processing, is verified with real coordinates. The experiments show target recognition with image processing can effectively improve the flexibility of our robot.