Study Of The Measurement Methods For Complex Surface In Reverse Engineering

With the development of measurement, Computer Aided Manufacture (CAM), Rapid Prototyping Manufacture and Rapid Tooling technologies, Reverse engineering (RE) has been a hot in the CAD/CAM field. In geometric reverse engineering, all kinds of measure technologies and apparatuses are used to get digital information from models firstly, and then we can remodel the measurand by some reverse software, like Surfacer, Imageware, etc., and finally the design drawing of the measurand is got. One important part of reverse engineering is the measurement of complex surface, which directly effects on the continent procedure of RE and the quality of final model. So study on it is necessary.The major work of this paper consists of measure programming of complex surface, adaptive tracing algorithm and surface fitting technology using Artificial Neural Network (ANN). There are six chapters altogether in this paper.In chapter one, the conception, flow and application of RE are formulated. Then the measurement devices and technologies used in RE are represented. The measurement technologies are divided into two classes, contact measurement and non-contact measurement. The characters and applications of each class are given.Then some adaptive tracing algorithms are analyzed to find their merits and demerits. And at last, the tasks of this paper are given.In chapter two, complex surfaces were divided into two classes by their features, regular surfaces and free-formed surfaces. Then the measurement programming for regular surfaces and free-formed surfaces was individually illustrated by examples.In chapter three, the adaptive tracing algorithm is studied. This chapter consists of five points as followed: theoretical basis of arc tangent extrapolation, principle of arc tangent extrapolation algorithm, arc length error evaluation of the algorithm, software flow of the algorithm, and experiments on the PMC650 comparing with the uniform sampling.In chapter four, the methods of parameter curve and surface fitting, major on Bezier, B-spline and NURBS, were introduced firstly. Then the surface fitting method-using ANN is advanced and examples are followed. The influences of network parameters to surface fitting are analyzed.In chapter five, the possible error sources of surface measure with CMM are analyzed and methods for compensation to temperature error and gauge head's radius error are illustrated.In chapter six, the conclusion of this paper and the expectation of reverse engineering development were declared.Many experiments on CMM PMC650 proved the algorithms mentioned in this paper were feasible, high efficient and easily realized. The digitalized description to the measurand was detailed, which facilitated later processes.