刀具旋转状态下切削刃的视觉跟踪及在线逐齿磨损检测方法

Online tool wear inspection for rotatable cutting tools such as milling tools is of great significance. Existing inspection methodologies are not perfect. The indirect measurement method can be applied for real-time inspection during machining process. However, there is room for improvement in the sensitivity and universalization of this mehtod. The optical image based direct method has the advantages of high accuracy and good universalization. But it has to stop the spindle and inspect tool wear condition of rotatable cutting tool manually. This research presents the new idea of combing 3D measurement of rotatable cutters, video motion analysis with optical image method for tool wear inspection. The theory and method to locate the cutting edge through video analysis and online tool wear inspection tooth by tooth are studied. The main research contents include: (1) The single view 3D reconstruction algorithm for surface of revolution is presented to solve the problem of 3D measurement of the sweeping body of a cutting tool. (2) The curvature segmentation of the generating function of a cutter is researched by the method of straight line approximation and combination for discrete second-order curve. (3) The optical flow method based on geometrical and kinematical constrain are presented to deal with the problem of cutting edge identification and tracking of a rotatable cutter. (4) The motion blur restoration and tool wear region segmentation are studied. The research and its achievement contributes to solve the problem of orientation of cutting edge of a rotating cutter and provide the methodological and theoretical basis for the application of online automatic tool wear inspection tooth by tooth for rotatable cutting tools.

在线监测铣刀等回转体刀具的磨损状态具有重要意义，现有检测方法尚不完善，间接测量法能够在加工中实时检测刀具磨损，但检测的灵敏度和通用性有待改进，直接测量法中的光学图像检测具有精度高和通用性好的优点，但需停机检测和人工辅助。本项目提出把回转体刀具的三维测量、视频运动分析与磨损检测的光学图像法相结合的新思路，研究在刀具旋转状态下通过视频分析对切削刃进行定位和在线逐齿磨损检测的理论与方法。主要研究内容包括：（1）采用回转体单视图三维视觉重建算法研究刀具扫掠体的三维测量问题；（2）通过离散二次曲线的直线逼近和合并方法研究刀具母线的曲率分割问题；（3）采用基于几何和运动约束的局部光流法研究回转体刀具切削刃的视频识别和跟踪问题；（4）研究磨损图像的运动模糊复原和磨损区图像分割方法。本研究及成果有助于解决旋转状态下刀具切削刃的定位问题，为回转体刀具在线自动逐齿磨损检测的应用提供方法和理论。

在线监测铣刀等回转体刀具的磨损状态具有重要意义，现有的检测方法尚不完善，间接测量法能够在加工中实时监测刀具磨损，但检测的灵敏度和通用性有待改善，直接测量法中的光学图像检测具有精度高和通用性好的优点，但由于回转体刀具有转动自由度，在机检测时切削刃的圆周角度和位置在显微视场中是不确定的，因此需要主轴停转并通过人工旋转和平移刀具，把切削刃调整到显微视场中合适位置，才能获取磨损区图像，这造成生产停顿、大大降低了生产效率。以上问题大大限制了光学图像法在自动化数控加工过程的实际应用。.本项目提出把回转体刀具的三维测量、视频运动分析与磨损检测的光学图像法相结合的新思路，研究在刀具旋转状态下通过视频分析对切削刃进行定位和在线逐齿磨损检测的理论与方法。主要研究内容包括：（1）采用回转体单视图三维视觉重建算法研究了刀具扫掠体的三维测量问题；（2）通过离散二次曲线的直线逼近和合并方法研究了刀具母线的曲率分割问题；（3）采用灰度响应法研究了回转体刀具切削刃的逐齿识别问题；（4）研究了基于直线检测的磨损区图像分割方法；（5）研究了景深扩展算法，克服了当前微细刀具在机检测中景深普遍不足和观测角度受限的问题。在对以上内容进行深入研究的基础上，先后开发了多套测量原型系统和刀具测量软件，通过实验验证了所提出方法的可行性。.本项目提出并实现了回转体刀具在机自动逐齿磨损检测的新方法，能够在主轴不停转的情况下，通过对旋转刀具图像序列的分析自动识别出每个切削刃的磨损图像，逐齿精准地分析刀具磨损情况。本项目提出的方法克服了光学图像法无法对铣刀等回转体刀具进行在机自动测量的局限性，为充分发挥光学图像法的优点提供了坚实的理论基础和技术支撑，将推动基于光学图像的刀具磨损在机检测技术与数控加工过程的深度融合。