Implementation of an automated real-time statistical process controller

Through a simple application, this research demonstrates the potential and procedure of applying computer vision and control techniques in developing automated real-time statistical process controllers (SPC). A computer vision system was developed to sample and measure quality variables (area, length, and width) of an extruded food product. Shewhart control charts were used to monitor the state of statistical control. Corrective actions were determined by using a PI control algorithm to minimize the product size variation caused by material inconsistency in moisture content. The system was implemented and tested. Considerable improvement was achieved in product size uniformity.