High Quality Sampling And B-spline Fitting Of Planar Curves With Feature Recognition

The high quality fitting method of curves and surfaces is an important research problem in computer aided design(CAD)and other related applications.For a parametric curve,the selection of sampling points plays an important role in the fitting results.Although the traditional uniform sampling method is relatively easy to implement,its sampling results often ignore the feature points that have important influence on the shape,resulting in poor fitting results.Therefore,the sampling method needs to select representative sampling points on the parametric curve to better grasp the contour features of the original curve.Aiming at the sampling problem of planar parameter curve,this paper proposes a high quality sampling method based on feature recognition.First,using parabolic interpolation to identify all the characteristic points on the target curve(inflection point and extreme curvature points)is simple and efficient.Then,the eigenfunction of the weighted combination of arc length and curvature is constructed,and more auxiliary points are adaptively selected by the eigenfunction,so as to obtain the sampling points of the given total number.According to the actual needs,the weights in the feature function can be further adjusted to provide greater flexibility and controllability for the selection of auxiliary points.Finally,the B-spline fitting curves of interpolating all feature points and auxiliary points are obtained by using the progressive iteration approximation.A large number of comparison examples show that this method has obvious advantages in sampling quality,efficiency and applicability.