Research On Off-line Extraction And Recognition Of The Form Information From Engineering Drawings For PDM

With the advent of AutoCAD, the computer-aided design has been deeply involved in the fields of machinery, electricity, building, clothing, etc. AutoCAD has become one of the most popular designing software in the CAD family. When we are designing the products based on AutoCAD, there used to be a lot of DWG files following.The drawing database management of PDM Set includes the capabilities of the database,network communications and the ability to control the process, which implement the exchange and sharing of information in the distributed environment and adjust and monitor the whole design process. It is necessary to extract information from the title and bill of material in the PDM system. Depending on database system, the management of the title and bill of material information was realized, which ensure that all datum is intuitive, orderly and controllable. There is a large number of CAD drawings document before implementation of product data management in enterprises. Before inputting manually information to the database on online, some useful drawing information should be stored into the database. By doing so, leaks and wrong items will appear inevitably. Moreover, when the drawing's titles and the contents of bill have changed, there is no corresponding changes on the information belonged to database. Thus, to some extent, it must bring about some inconvenience to designers.Therefore, to avoiding the tedious process of the resume load, information should be extracted by off-line methods to analyze and identify. The automatic extraction information becomes especially important in order to making PDM systems to disengage CAD software by interface technology.In view of the present situation, this article has carried on the comprehensive analysis to some existed achievements. In this paper, the computer graphics,pattern recognition and CAD re-development theoretical knowledge were referred, researching on off-line recognition and the extraction of the form information system of the algorithm and implementation. The method propsed in this paper has been applied in a project of Anhui Jianghuai Special Purpose Vehicle Co.Ltd, named "Variety and Small quantities Special Vehicle Product Data Management System". Experimental results show better processing qualities. The function putted forward is to automatically extract the information from form, and save to the database, avoiding the manual work and enhancing the accuracy.