面向STL模型的在机检测系统测点法矢计算
Calculating Normal Vector of Measurement Point for STL Model in On-machine Measurement System

为提高在机检测系统的测头半径误差补偿精度，针对测头半径误差补偿的重要参数-测点的法矢方向提出了精确的估算方法。针对测点位于三角网格顶点的情况，提出一种以三角面片的顶角、质心距及周长作为权重因子的顶点法矢的估算方法。采用三种二次曲面的三角网格模型进行实验验证，结果表明，针对测点位于三角网格顶点情况的计算方法可有效地提高法矢计算精度。同时，针对采用截面线法规划测点时，测点普遍位于三角网格内部的情况，提出了计算测点的法矢方向，并通过两种不同网格密度的正弦曲面进行实例验证。结果表明，采用面向STL模型的在机检测系统测点法矢计算方法计算的测点法矢误差的平均值较其他方法分别减少了53.5%、61.7%，使测点法矢的计算精度明显提高，从而有效地提高了半径误差补偿的精度。
To improve the radius error compensation accuracy of an on-machine measurement system, an accurate estimation algorithm is presented for the normal vector of a measurement point, which is an important parameter for radius error compensation. When the measurement point is in the vertices of triangular meshes, a normal vector estimation method is proposed based on the angle, centroids distance and perimeter of a triangle and is validated by the triangular mesh models of three quadric surfaces. The experimental results demonstrate that the improved algorithm can effectively improve the accuracy of normal vector estimation. Meanwhile, under the condition that the measurement points planned by the cross-section curve method are located in the interiors of the triangle, the barycenter coordinate method is proposed to compute the normal vectors of measurement points and verified by the density of two different grids with sine-like surfaces. The experimental results show that the improved algorithm decreases the average normal vector error by 53.5% and 61.7% respectively and obviously enhances the accuracy of normal vector estimation and radius error compensation