B样条曲线参数重载的非线性整体逼近拟合法

Non -linear curve approximation using integral b -splines

论述了一种新的自由曲线的非线性拟合算法。此算法应用最小二乘法来获得最佳拟合曲线。解析和数值积分法的结合使用实现了曲线整体逼近,从而避免了对原曲线采点取样。B样条的曲线表达使之适用于各种复杂曲线的拟合。此算法结合渐进式逼近法可满足不同拟合精度要求。相对于传统的采点插值,或取样近似法,此方法更简洁、快速,拟合效果更佳。

Outlines an algorithm for non -linea r approximation of procedurally defined curves.The algorithm uses the least square meth od to minimize the integral error,which eliminates the need for sam ling t he ideal curve.Ahigh -level iterative procedure allows one to approximate a variety of curves and ensures that t he deviations between the ideal and app roximated curves are within user -sp ecified error limits.The proposed scheme is more efficient and robust t han traditional interpolation and a pproximation methods.