提出了一种新的轮廓滤波器,它以高精度逼近高斯滤波器.这种轮廓滤波器是通过B样条与变分法相结合构造出来.通过调节逼近参数,该滤波器与高斯滤波器的逼近偏差可以控制在0.94%.该滤波器的算法是在频域中实现而不是时域中,所以大大缩减了计算时间.该滤波器的另外一个重要特性是有效地抑制了出现在滤波结果数据边缘的失真,也即边缘效应,这是许多轮廓滤波器所共有的缺陷.对一个实际工件进行测量提取表面中线和粗糙度,结果令人满意. A new contour filter is proposed, which approximates the Gaussian filter with high accuracy. This contour filter is constructed by combining B-spline and variational method. By adjusting the approximation parameters, this filter is similar to the Gaussian filter The approximation bias can be controlled at 0.94%. The filter algorithm is implemented in the frequency domain rather than in the time domain, thus greatly reducing the computational time. Another important feature of the filter is to effectively suppress the filter data The distortion of the edge, ie the edge effect, is a common defect of many contour filters. The measurement of an actual workpiece to extract the surface centerline and roughness gives satisfactory results.