A generic dynamic inspection system for visual object inspection and industrial quality control

This paper describes the development and application of a hybrid and dynamic system concept for visual object inspection in industrial quality control. This system incorporates image processing techniques, knowledge processing, and neural as well as non-neural methods for classification and data analysis. These modules are not hardwired in a processing chain, but can be flexibly configured under the control of a central unit that surveys the inspection process. The system assumes a configuration based on a priori knowledge and on the results of the self-monitoring process. In this work the first experiments and results utilizing a prototype implementation of this concept are presented.