Title :
CELESSTIN IV: knowledge-based analysis of mechanical engineering drawings
Author :
Vaxivière, Pascal ; Tombre, Karl
Author_Institution :
ESSTIN-CRIN/CNRS, Vandoeuvre, France
Abstract :
In automated analysis of technical documents, the main objective is to be able to retrieve the information described by paper documents in order to convert it to a format suitable for computer-aided design (CAD) systems. It is shown how high-level knowledge can be used to analyze an engineering drawing from a functional point of view. The authors present the knowledge rules used in the CELESSTIN IV system: technologically significant entities are extracted and the whole setup is analyzed with respect to disassembling and kinematic knowledge. CELESSTIN IV is only a prototype system, but an integrated one, providing in the same software package all the necessary tools: image preprocessing, vectorization, thin and thick lines separation, layer segmentation, analysis of dimensions and optical character recognition (OCR) of the text parts, decomposition of the drawing into blocks, labeling of the blocks along axis lines, and higher-level knowledge-based analysis of the drawing in functional terms
Keywords :
engineering graphics; image processing; image segmentation; knowledge based systems; CAD; CELESSTIN IV; image preprocessing; knowledge-based analysis; labeling; layer segmentation; lines separation; mechanical engineering drawings; vectorization; Design automation; Engineering drawings; Image analysis; Information analysis; Information retrieval; Kinematics; Mechanical engineering; Optical character recognition software; Software prototyping; Text analysis;
Conference_Titel :
Systems Engineering, 1992., IEEE International Conference on
Conference_Location :
Kobe
Print_ISBN :
0-7803-0734-8
DOI :
10.1109/ICSYSE.1992.236861