截面测量数据的B样条曲面拟合

Bicubic B-spline surface fitting based on data from section measurement

针对激光扫描仪测量得到的截面型数据,提出了对截面线重新采样的方法,分析了重采样数据拟合曲线的误差,及平方误差随重采样比例变化的规律,从而通过控制误差来确定重采样数目。构造插值于截面数据点的非均匀B样条,再对其进行相同数目的采样,得到了呈矩形分布的点阵,采用双向平均积累弦长参数化方法确定、方向节点矢量,用插值的方法进行曲面拟合。此种方法不仅避免了对轮廓线进行相容性的处理,而且减少了模型数据量,提高了曲面重构的效率。

According to the cloudy data from section measurement by laser scanner, a method of resampling the B-spline section line was proposed. Resampling data fitting error was provided in this paper, and the rational number of sampling point can be determined according to the sampling accuracy required. The resampled data distributes in rectangular array by sampling the same number in each line. knot vectors for constructing the longitudinal and transverse boundaries are defined as a compromised average of all section curves in each direction respectively.Non-uniform B-spline surface are exploited to fit obtained resampled data. This method make sure knot vectors of contour line the same not by inserting knot, which not only reduce the original data but also increase surface reconstruction effwiency.