Reverse Engineering Feature Extraction Techniques

Combined with the rapid development of laser technology and numerical control technology, Reverse Engineering is becoming a significant product design method in the industrial fields of auto, aviation, mould, toys and shipping. Reverse Engineering obtains structure data and model parameters based on substantial model, absorbs advanced designing and building ideas, even management methods during manufacture by model restructure, leads to products that gear to the market more properly by modification and redesign. The designing idea & building technology ,which dramatically reduce production cycle and increase productivity, is a kind of CAD technologies involved into many subjects such as computer graphics, differential geometry, differential geometry, mathematics statistics, software engineering and so on.Ant colony algorithm devised by M.Dorigo and his co-workers from Italy is a new developing simulation optimization algorithm based on ants seeking food. With the positive feedback parallel autocatalysis system Ant colony algorithm hold the meits such as forceful robustness, advanced distributing compute system, cooperating with other methods easily and so on and have displayed its predominance and developing potential.In this paper, the ant colony algorithm (ACA) was first applied in reverse engineering for data processing, and solved the problems of feature extraction, data compositor and data optimization successfully. The main research work can be described as follows:A method was given to extract the feature of the cylinder surface in discretionary 3D space to solve the problem of coordinating in the measurement technology. The extraction principle was demonstrated and its correctness was verified by some examples.The extraction principle of the revolving surface in discretionary 3D-space was presented and the symmetry axis and its vertical of the section data based on the infinitesimal inertia principle were carried out. Finally, experimental results were presented to verify the feasibility of the new algorithm.The ant colony algorithm was proposed for feature recognition of revolving surface and it was the first time that the ant colony algorithm (ACA) has been applied in reverse engineering for data processing. The mathematic model was established which was suitable for extraction of feature parameter using ACA. The example testified the accuracy and efficiency of the method.The data points of the section curve were out-of-order, so the TSP principle of ant colony algorithm was applied to give the right sequence in this paper and it did not only deal with the 2D datum but the points in 3D-space. The results of experiment indicated that the taxis were accomplished simply after one iteration.Finally, the paper introduced the reconstruction of NURBS surface. The section curves were extracted by delaminating the point cloud and fitting the section points, and the NURBS surface was reconstructed based on the section curve in Pro/E system. No-surely minimum distance data compression method and no-surely equivalent depth delaminating method were applied in paper. Compared to the usually methods, the two methods are not only simple but also can keep the detail features of the point cloud.