KNOWLEDGE BASED REVERSE ENGINEERING METHODOLOGY

摘要

In Reverse Engineering, commercial solutions such as Geomagic™ or RapidForm™ or CAD software such as CATIA™ provide very efficient toolboxes that enable to rebuild geometry. Moreover, some of them provide segmentation algorithms, sketchers and/or many other facilities rebuilding tools in order to help to recover the original surface. Those software's enable to rebuild the geometry of a technical object as a set of functions (protrusion, revolution, sweep...), they enable to add colours and textures, and they enable realistic kinematical animation and many other things. Unfortunately, all those toolboxes present a lack of geometry analysis tools. The geometry of a given product is the consequence of a complete process, it is important, considering Reverse Engineering activities, to try to recover any evidence of its past life (including socio-economical aspects, the design intents of the former designer, its different uses...) from its geometry in order to produce a good quality 3D model. Such models can provide important possibilities for Reverse Engineering. It enables to study the product more efficiently than a geometrical model based on a mesh or on free form surfaces. This paper proposes approaches and methodologies for using knowledge to rebuild CAD model to the best closer to the original CAD model that could have been obtained by the original designer. Such a model is feature based. They can be functional features or manufacturing features. For a good rebuilt of 3D model, the geometry of these features is driven by parameters, rules and relationships that are provided by the former designer.