Indexed on: 23 Dec '18Published on: 19 Dec '18Published in: International Journal of Engineering & Technology
The best method in developing automated assembly system is the integration of the machine vision tools with the robot in the assembly platform. Vision system plays an important role in building an autonomous robotic part assembly system. Recognition of the correct part in the assembly line is the key issue while grasping the parts by the robot. The captured images of the parts on the assembly line may be affected by geometric transformation such as rotation, translation, scaling and may be corrupted by Point-Spread-Function (PSF) blurring of camera. In order to recognize parts in such type of condition, three steps are followed in this paper. Firstly, features of the parts are extracted by using combined orthogonal Zernike moment. Secondly, the original part image is reconstructed by using combined orthogonal Zernike moment. For this reconstruction, the moment order is selected by a proposed algorithm. The optimum moment order will maintain a proper balance between the reconstruction capability and sensitivity to noise. Finally, for classification of parts used in assembly nearest-neighbor classifier is used. The suggested technique is implemented in LabVIEW and the simulation is successfully performed in an assembly system with 6-DoF Kawasaki robot.