The presence of residual burrs on the surface of workpieces not only affects the surface quality but also reduces their performance. The position and orientation of the cutting tool in the deburring process are crucial for effective burr removal. Due to the interference caused by the burrs uncertainty or the manufacturing deformation, the off-line path from the workpiece CAD model is difficult to ensure the processing quality in robotic deburring. To solve that, a correction strategy for the robotic deburring paths is developed based on a visual feedback model registration. Firstly, by applying CAM software, the nominal tool path is generated from the workpiece CAD model. Secondly, by conducting global rigid registration, followed by a local non-rigid registration with burr height constraint, the transformation matrix between the measuring data and CAD model can precisely calculate to mitigate the impact of local shape deformations on registration precision. Finally, the actual tool path is obtained by using the transformation matrix to correct the nominal tool path. Comparisons results show that: for the proposed method, the maximum contour deviation between the corrected path and reconstructed theoretical is less than , while the chamfer size remains relatively uniform in real machining with minimal variation ranging from to , which demonstrates that such method can effectively address the issue of deviation in robot deburring paths caused by workpiece deformation.

中文翻译：

---

使用局部非刚性 3D 配准进行机器人去毛刺的刀具路径校正


工件表面残留毛刺的存在，不仅影响表面质量，而且降低其性能。去毛刺过程中切削刀具的位置和方向对于有效去除毛刺至关重要。由于毛刺不确定性或制造变形造成的干扰，工件CAD模型的离线路径很难保证机器人去毛刺的加工质量。为了解决这个问题，基于视觉反馈模型注册开发了机器人去毛刺路径的校正策略。首先，通过应用CAM软件，从工件CAD模型生成标称刀具路径。其次，通过进行全局刚性配准，然后进行毛刺高度约束的局部非刚性配准，可以精确计算测量数据与CAD模型之间的变换矩阵，以减轻局部形状变形对配准精度的影响。最后利用变换矩阵修正名义刀具路径，得到实际刀具路径。对比结果表明：对于所提出的方法，校正路径与重构理论之间的最大轮廓偏差小于 ，而实际加工中倒角尺寸保持相对均匀，变化最小范围为 到 ，这表明该方法可以有效解决工件变形导致机器人去毛刺路径偏差的问题。