The Research On An Example-Driven Symbol Recognition Algorithm Based On Key Features In Vector Drawings

With the development of CAD software platform,2D digital drawings have becomethe communication tool in the design and engineering feld. How to understand anduse the information behind the drawings is the hot issue in recent year. There are a lotof symbols, which indicate the semantic message of the drawing. So it is important torecognize the symbol from the plan efciently and correctly.In this paper, an example-driven symbol recognition algorithm based on key featuresis proposed for CAD2D drawings. Firstly, we extend the existing relation representationby involving new basic elements and redefning the relation attributes between them. Foran input symbol, its structure can be analyzed and represented by the relation mechanismautomatically in the form of a graph. Then, based on the relation graph, the constrainedtree with key feature priority can be established for this type of symbol using the knowl-edge acquisition algorithm. The constrained tree is considered as the recognition rule andstored in the symbol library. In this way, it is convenient and efcient to build and expandsymbol libraries to handle various engineering drawings. In the recognition process, thekey feature node can be located frstly in the drawing. The other elements around, whichsatisfy the topology structure of the constrained tree, will be found next. The symbolobject will be recognized if all the elements and constrains in the tree are found. Theunnecessary matching calculations would be greatly reduced because of the accuratelypositioning. Moreover, an extra pretreatment analyzing approach is proposed in order todeal with the repeat mode in a symbol. Thus, similar symbols can be recognized by onlyone rule. Experimental results validate that the approach is efective.