Research On The Composite Freeform Surface Models Based On Adaptive Accuracy Assessment

In China, 65% of the development and introduction of technology in recent decades is application in reverse engineering, and the precision reverse of a part which has composite free-form surface features is challenge in this field. In order to achieve high efficiency and high precision of reverse, the evaluation on accuracy of surface model in real-time is taked in the traditional reverse to form a closed loop process.On the basis of deep study on existing measurement techniques, Coordinate Measuring Machine(CMM) is used in order to obtain accurate measurement data. Before measurement, planning of decompose complex free-form surfaces and free patch measuring order is needed. Two-parameter adaptive measurement method is applicated to each free patch. After the initial test in U parameter direction, fitting the measurment points to a high order B-spline curves, predicting the next measuring point location according to the curvature of the fitting curves. In V parameter direction, fitting the measured point by using B-spline surface. The next scan line is adaptively determined based on the surface of the maximum curvature of V direction. The measurment points can be distributed due to the characteristic changes of the surface, the points are intensive in the important features area, and sparse in unimportant area. As a sequence, the important features points will not be missing and data redundancy is avoided. Some example applications show that the measurement accuracy up to μm level.Because of the characteristics of adaptive distribution of measurement points density, surface reconstruction based on the non-uniform B-spline surface interpolation is proposed. Through reconstruction experimental of adaptive measurement points of saddle, the results show that the maximum deviation is less than 6 μm between the surface model and its mathematical model structure, the average deviation is about 1-2 μm. The efficiency is 10 times faster than Imageware manual reconstruction.To overcome the traditional reverse defects and judge the differences between the model and the part immediately, an adaptive method for evaluation of model accuracy is proposed. Some points are choosen as the evaluation points in real-time based on the reconstruction model’s control mesh, rational planning of the evaluation points measurement path directs output of DMIS program, and guides CMM automatically to measure these points. The distance between the evaluation points and measurement points is calculated, and the model error is evaluated. When unqualified, the measurement points are added into points cloud with reconstruction and accuracy assessment again, the times of evaluation is adaptively determined by the model itself error.Using VC++, Open GL, LK-DMIS, CMM and other hardware and software, closed-loop reverse system is developed, including an adaptive measurement, reconstruction, and adaptive accuracy assessment. Through a large number of examples, results show that this system can reduce the reverse cycle and cost.