Panter — Knowledge based image analysis system for workpiece recognition

Abstract

This article describes a knowledge based image analysis system that uses the Hierarchical Structure Code (HSC) for the initial segmentation of the image data. The generation of this code is domain independent and allows for an elegant coupling with a knowledge base. The system was successfully tested in the field of workpiece recognition. In the first section of the article, a short introduction of the image transformation is presented, a presentation of methods to extract the size and position invariant features follows. The main emphasis of the article is on the description of a semantic network language for a knowledge based image analysis system using the Hierarchical Structure Code.