Intelligent system for inspection and selection of parts in a manufacturing cell

This paper addresses the design and implementation of an artificial vision system implemented in a manufacturing cell. The vision system recognizes and selects in an intelligent manner the manufactured parts through a feedforward artificial neural network and the decisions are completely based on the part's color and its geometry. A simple digital camera is used as an image acquisition device. This image is then processed by an artificial neural network, which is able to identify the part's color. Then a Programmable Logic Controller (PLC) drives an electropneumatic system, in order to store the identified part into a corresponding repository. An interface based on power electronic devices and a Data Acquisition Card (DAQ) system implements the communication between the PLC and the computer. The proposed system is completely implemented and tested in a real Flexible Manufacturing System (FMS) of FESTO© showing good results.