Create A Distributed Library And Shape Matching Methods

Reuse of drawings can insure the criterion of design, improve the efficiency of design, and shorten the period of the product development. Sketch-based retrieval uses designer's rough drawings to search the objective from existent drawings or drawing database and it is one of the objects that design reuse of CAD are always looking forward to. The basic of sketch-based retrieval is graph matching. At present, the mostly used method of graph matching is to use the graphic feature and graphic syntax, while the result is not perfect. This paper limits to planar mechanical CAD drawings and aims to establish a new method for sketch-based retrieval.This paper analyses problems appeared in graph matching, utilizes the theories of Fuzzy Mathematics and Computer Graphics, then puts forward the direction of graph matching. It's difficult to compare two drawings' similarity using conventional CAD graphic model. Thus an automatic logic model is reconstructed based on graphic feature expression. The new model uses feature nodes as an element to describe graph's topological structure and geometric characteristics, and then forms a sort of comparable essential cell for graphs. Simultaneously, a method of graph matching is presented via essential cell compare. It indicates that the method has highly accurate rate and very good adaptability in the application. Furthermore, considering the requirement of the practical item, this paper studies the distributed drawing database, carries out the retrieval system of the dockyard center graphic database and also implements the client subsystem, graphic retrieval subsystem and synchronization system to keep the consistence of center and client graphic databases..The rebuilding of graphic modal well resolves the matching of planar mechanical CAD drawings and obviously it is a profitable attempt. This new idea can also be applied to other areas of Computer Graphics. Distributed construction improves the speed of searching drawings and it is a very good technology for projects with large amounts of data to resolve problems effectively. Both the retrieval method and the retrieval system that supplied in this paper received good evaluation from the designers and make the original aims come true.