Three approaches for the detection of CAD-part attributes as a preparation for an automatic classification

In this paper we present a research project for the automation of a classification for machine parts. We developed algorithms for the automatic detection of attributes of machine parts from neutral CAD-file formats. Here three different approaches are presented. The first deals with the reconstruction of CAD-models based on images or freehand sketches. The second one analyses the construction chronic of the parts aided by XML-files and the last approach investigates in the neutral file format of STL-data. Computer programs based on these algorithms were tested with a collection of CAD-parts. We describe the basic assumptions and the mode of operation of these algorithms and show first results of this research project.