Research On Inspecting Feature Extraction And Recognition Of Sheet Metal Parts Based On Three-Dimensional Model

In order to make a reasonable inspection planning for the aircraft sheet metal parts,whose structure is complex; we must extract and recognize the inspection characteristicsof the sheet metal parts from the three-dimensional model of the sheet metal parts. Theinspecting characteristics recognition of sheet metal parts is that, we should extractgeometry information with specific engineering significance from the3D design ofsheet metal parts model, and then it can generate a feature model of product inspection.The main research content of this thesis is about that how to extract inspectioncharacteristics of sheet metal parts from the3D model of sheet metal parts. In order tomeet the target, this thesis puts forward two kinds of treatment methods. The firstmethod is used for the3D design model of general aircraft sheet metal parts which isdesigned in the CATIA, using the combining of the CATIA secondary development andthe Drools rule-engine to get the inspection characteristics from the3D CATIA designmodel. The second method is about the general3D model and some characteristicswhich are difficult to describe in the rules file. Considering this situation, the thesispicks a processing format based on the STL files to get the inspection feature of thesheet metal parts extracted and identified. Therefore, the main content of this thesis canbe summarized as follows:First of all, this thesis researches on the CATIA which is commonly used modelingsoftware of the sheet metal parts. And then, the thesis carries on the secondarydevelopment and extracts the feature information related to the inspection of sheet metalparts.Secondly, this thesis researches the Rete algorithm and Leaps algorithm deeply. Onthe basis of the Drools rule engine, the knowledge base of sheet metal parts inspectioncharacteristics is established, according to the characteristics of sheet metal parts inspection specifications sheet metal parts in3D model and the actual characteristics ofthe specific form of expression.Finally, in virtue of that the STL file is a commonly used method to store anddeliver3D-model-data, this thesis studies the theory of region segmentation algorithmbased on graph-based feature recognition emphatically. And finally, this thesis picks updifferent characteristics of recognition algorithm to achieve the identification of sheetmetal parts inspection characteristics, according to the characteristics of different types.