Annotator: an AI approach to engineering drawing annotation

Abstract

I. AbstractAnnotator is a prototype to investigate the application of AI techniques to the annotation of engineering drawings. In particular, Annotator addresses drawings of piping systems such as those for chemical plants or waste treatment facilities. The isometric representation of the piping system is selected because it is the most numerous type of drawing in plant design. Knowledge contained in hierarchies represents the CAD model of the piping system, features of the model and features of the drawing. Production rules about relationships among model items and features make annotation decisions, creating instances in a hierarchy of drawing annotations. Issues in geometry such as parallelism and perpendicularity and the translation of the 3-D model world into the 2-D drawing representation are addressed. Algorithms determine geometric intersections; rules specify whether intersections between given object classes are considered acceptable. Although Annotator prototype is separate from the CAD/CAM environment, the Annotator prototype illustrates the potential for fully automating the annotation process through representation of the knowledge about a piping model and its dimensioning requirements within a rulebased, object-oriented hierarchy.

Knowledge Annotation: Making Implicit Knowledge Explicit

“…Artificial Intelligence approaches too help in the creation of vector annotations such as in [220]. In this application, they make use of different techniques to annotate piping systems (such as those in waste treatment facilities, chemical plants, etc).…”