Application of knowledge-based image inspection system for diagnosis of misprints in offsetprinting

Abstract

In the paper the development of a fault detection and diagnosis system for continuous quality monitoring and control in offsetprinting is described which was developed to support the operator in finding his decision during disturbances. Real-time capability was ensured by splitting the system into a high speed image processing unit and an expert system. The image processing unit takes the sheet, detects the misprints and extracts the features. These features are than transferred on-line to the expert system shell SWD-expert, which identifies the type of the misprints and gives control recommendation to the operator how to eliminate the disturbance. In the paper we describe the methodology for the development of such a system as well as the function and the developed algorithms of each processing stage.