Accurate gear profile plays an important role in determining the transmission performance of gear-drive equipment. In this paper, a novel method for extracting gear tooth profile edge is presented. The presented method is based on engagement-pixel image edge tracking (EPIET) technique, and does not rely on the traditional meshing theory. An algorithm for the proposed method is put forward. Firstly, instantaneous contact images between the envelope curves of the gear profile and the instantaneous locus of the cutting tool are acquired. Secondly, pixels on the boundary of the envelope curves are lighted and the instantaneous locus coordinates of the cutting tool are calibrated. Lastly, the pixel coordinates of instantaneous meshing points are extracted, based on the fact that there is exactly one contact point per instant, and major error sources of the presented method are discussed. To verify the effectiveness of the presented method, a case study is conducted on a face gear, which is a type of complex conjugate gear, to extract its tooth profile edge. In the study, the tooth profile error and the contact line error are investigated. The results demonstrate that the presented method, without knowing complicated meshing equations, can acquire feasible accuracy and stability, compared with traditional meshing equations. It is shown that the novel method can be extended to applications of digital design of complex conjugate curved surfaces, in a simple and fast manner.

中文翻译：

---

一种新的齿轮齿廓边缘提取的啮合像素图像边缘跟踪方法

准确的齿廓在决定齿轮传动设备的传动性能方面起着重要作用。在本文中，提出了一种提取齿轮齿廓边缘的新方法。所提出的方法基于接合像素图像边缘跟踪（EPIET）技术，不依赖于传统的网格划分理论。提出了所提出方法的算法。首先，获取齿轮轮廓包络线与刀具瞬时轨迹之间的瞬时接触图像。其次，点亮包络线边界上的像素，校准刀具的瞬时轨迹坐标。最后，基于每一时刻恰好有一个接触点这一事实，提取​​瞬时网格点的像素坐标，并讨论了所提出方法的主要误差来源。为了验证所提出方法的有效性，对面齿轮（一种复共轭齿轮）进行了案例研究，以提取其齿廓边缘。在研究中，研究了齿形误差和接触线误差。结果表明，与传统的网格划分方程相比，所提出的方法在不知道复杂网格划分方程的情况下，可以获得可行的精度和稳定性。结果表明，该新方法可以以简单快速的方式扩展到复共轭曲面的数字设计应用。提取其齿廓边缘。在研究中，研究了齿形误差和接触线误差。结果表明，与传统的网格划分方程相比，所提出的方法在不知道复杂网格划分方程的情况下，可以获得可行的精度和稳定性。结果表明，该新方法可以以简单快速的方式扩展到复共轭曲面的数字设计应用。提取其齿廓边缘。在研究中，研究了齿形误差和接触线误差。结果表明，与传统的网格划分方程相比，所提出的方法在不知道复杂网格划分方程的情况下，可以获得可行的精度和稳定性。结果表明，该新方法可以以简单快速的方式扩展到复共轭曲面的数字设计应用。