The cutting edge radius is a significant parameter of a micromachining tool. When it is not sufficiently sharp, ploughing, which affects the whole machining process and workpiece quality, will occur. It is then essential to be able to estimate the value of the cutting edge radius accurately. In this paper, a three-dimensional strategy that can be used to measure the microturning tool edge radii is presented. The strategy is based on robust cylinder fitting that is applied to a 3D point cloud of the tool. It is implemented in C++ with Open Computer Vision and Point Cloud Library. It was validated using a virtual point cloud, resulting in errors of 0.42% and 1.11% without noise and with noise, respectively. Additionally, uncertainties of m and m were obtained. It was successfully applied to six microturning inserts: three unused tools and three used tools. The point clouds were obtained with two different 3D surface reconstruction techniques, focus variation with a photon microscope and a multi-view stereo with a scanning electron microscope. The obtained results were more coherent, i.e., they were less dispersed; for example, in non-used tools, the range was [m–m] and for the conventional circle fitting the range was [m–m]. The traditional method is indirect: the 3D point cloud is sliced into 2D point clouds (profiles) fitted with circles. This 3D-2D process might result in errors. The proposed method is a direct 3D approach with no slicing step.

中文翻译：

---

微车刀刃口半径的三维测量

切削刃半径是微细加工刀具的一个重要参数。当不够锋利时，就会出现犁刀现象，影响整个加工过程和工件质量。因此，必须能够准确估计切削刃半径的值。本文提出了一种可用于测量微车刀刃半径的三维策略。该策略基于应用于工具 3D 点云的稳健圆柱拟合。它是用 C++ 实现的，具有开放计算机视觉和点云库。使用虚拟点云进行验证，无噪声和有噪声时的误差分别为 0.42% 和 1.11%。此外，还获得了 m 和 m 的不确定度。它已成功应用于六个微车削刀片：三个未使用的刀具和三个使用过的刀具。点云是通过两种不同的 3D 表面重建技术获得的，即使用光子显微镜进行焦点变化和使用扫描电子显微镜进行多视图立体。获得的结果更加连贯，即分散性较小；例如，在未使用的工具中，范围为 [m–m]，对于传统的圆拟合，范围为 [m–m]。传统方法是间接的：将 3D 点云切片为配有圆圈的 2D 点云（轮廓）。此 3D-2D 过程可能会导致错误。所提出的方法是一种直接 3D 方法，没有切片步骤。