A New Reverse Engineering Methodology Based on Knowledge Extraction and Geometrical Recognition Approaches

Reverse engineering (RE) is a research field where physical models are measured or digitized in order to be virtually rebuilt to obtain a CAD model. Today, rebuilt CAD models using current geometrical approaches are generally “frozen” or barely re-usable. This paper deals with a new reverse engineering approach. This one improves the classical geometrical approach with a knowledge extraction phase. This improvement enables to define a new methodology that assists the user during the reverse engineering process. The obtained CAD model is not a set of surface but a real set of design features structured in a design tree with constraints, parameters, rules and relationships. This structure enables to “unfreeze” the geometry of the model and really allows to re-study and to redesign the original physical model.