Welcome
Digital library of construction informatics
and information technology in civil engineering and construction
 

Works 

Paper 9687:
Converting CAD Drawings to Product Models

Facilitated by the SciX project

Noack, Robert

Converting CAD Drawings to Product Models

Abstract: The fundamental aim of this study is to examine whether it is possible to automatically convert vector-based drawings to product models. The reason for doing this is that the new object-based systems cannot make use of the information stored in CAD drawings, which limits the usability of these systems. Converting paper drawings to vector-format is used today and provides recognition of lines and text, but does not interpret what the shapes represent. A language for describing the geometrical representations that could be processed directly into a recognition program for building elements is missing. It is easier to describe how to recognize a line as a series of dots in a raster image, than it is to describe how a complex symbol of a building element looks like. The approach in this research work has been to test different shape recognition algorithms. The proposed method can be divided into four processes: grouping of geometrical primitives, classifying these groups, interpreting the content and analyzing the relationships between the groups. The algorithms developed here are based on research within related domains, such as pattern recognition and artificial intelligence. The algorithms have been developed in a prototype implementation and were tested with three layer-structured drawings used in practice. The results of the tests show that there are no crucial obstacles to recognizing a large part of the symbols of building elements in a CAD drawing. The requirement is that the recognition system is able to differentiate one from another and be tolerant of errors and variations in the shapes.

Keywords: Product Model, CAD

DOI:

Full text: content.pdf (1,170,022 bytes) (available to registered users only)

Series: other (browse)
Cluster:
Class:
Similar papers:
Sound: N/A.

 

hosted by University of Ljubljana



includes

W78




© itc.scix.net 2003
Home page of this database login Powered by SciX Open Publishing Services 1.002 February 16, 2003