Machine vision for feedback control in a steel rolling mill

The image processing algorithms use statistical estimates of confidence to ensure robust operation under extreme conditions. Feedback-control requires a sampling rate approximately 10 times higher that the bandwidth of the system which is to be controlled [K.J. Astrm, T. Hgglund, PID Controllers: Theory, Design, and Tuning, Instrumental Society of America, 1995 [1], Jukka Lieslehto, Wap application for pid controller tuning, in: Proceedings of the 2000 IEEE International Symposium on Computer-aided Control System Design, vol. September 25-27, Ancorage, Alaska, USA, 2000, pp. 168-172 [2]]. The system demonstrates the acquisition and processing of 500 frames per second in real time, this is commensurate with 50 Hz bandwidth presently possible for hydraulic systems. This opens the door to real-time feedback control using digital image processing.