Industrial signal processing

EvaSys: Zustandsbeurteilung komplexer Systeme zur Ableitung von Operationsempfehlungen

02.11.2004 bis 31.10.2005

The authenticity checking and inspection of bank notes is a high labour intensive process where traditionally every note on every sheet is inspected manually. However with the advent of more and more sophisticated security features, both visible and invisible, and the requirement of cost reduction in the printing process, it is clear that automation is required. As more and more print techniques and new security features will be established, total quality security, authenticity and bank note printing must be assured. Therefore, this factor necessitates amplification of a sensorial concept in general. We propose a concept for both authenticity checking and inspection methods for pattern recognition and classification for securities and banknotes, which is based on the concept of sensor fusion and fuzzy interpretation of data measures.  In the approach different methods of authenticity analysis and print flaw detection are combined, which can be used for vending or sorting machines, as well as for printing machines. Usually only the existence or appearance of colours and their textures are checked by cameras. Our method combines the visible camera images with IR-spectral sensitive sensors, acoustical and other measurements like temperature and pressure of printing machines.

This project is promoted by:
Real-time image-processing-system-on-chip for security feature detection and classification.
In: IST/SPIE 16th Annual Symposium on Electronic Imaging - Real-Time Imaging VIII, Jan 2004
A Simplified Scheme For Hardware-Based Pattern Recognition
In: IEEE International Conference On Image Processing (ICIP), Sep 2005