Syntactic interpretation approaches for innovative engineering drawings conversion support system

Abstract

By integrating human intelligence and enabling geometric modeling and knowledge based systems technologies, a realistic and plausible solution for two critical activities in a legacy data conversion cycle of engineering drawings (LDC): vector drawing editing and high level symbol interpretation was investigated. A prototype system called LEDCONS was implemented to illustrate the concepts and demonstrate the feasibility of the architecture and methodologies proposed. In this paper, the concepts and methods used by LEDCONS to perform syntactic level drawing interpretations are presented. The LEDCONS system modules at this level are designed based on the strategy that drawing interpretation could be embedded within the process of human editing in an interactive fashion. Two fundamental approaches, the automatic search-and-match approach and the manual select-and-recognize approach, used for recognizing drawing constructs at this level are proposed. With syntactically related drawing constructs, such as text blocks, hatch patterns, dimension sets, FCF/datum constructs, and geometry entities are extracted. Hence, the reusability and reliability of the drawing data can be significantly increased.