Abstract: In the wheel hub grinding and polishing robot system, due to the different shapes of the hub and the complex surface structure, the traditional contact sensor is not suitable for the alignment of the initial point of grinding and polishing. Therefore, a vision-based initial point alignment method of the hub is proposed. The entire initial point alignment process includes three stages：the first stage collects the image of the hub, and obtains the center coordinates of the hub and the center of the bolt hole. In the second stage, according to the center coordinates of multiple bolt holes and the center coordinates of the hub, the rotation angle required by the current wheel hub grinding and polishing point to the target initial point is calculated. In the third stage, based on the fuzzy control algorithm, the alignment controller is designed to realize the initial point alignment control. Finally, the effectiveness and accuracy of the proposed method are verified by experiments. The experimental results show that the method of aligning the initial point of wheel hub grinding and polishing is feasible.

Key words: machine vision, wheel hub grinding and polishing system, fuzzy control, initial point alignment

摘要： 在轮毂磨抛机器人系统中，由于轮毂形状各异且表面结构复杂，传统的接触式传感器不适用于磨抛初始点的对准。因此提出一种基于视觉的轮毂初始点对准方法。整个初始点对准过程包括三个阶段：第一阶段采集轮毂图像，得到轮毂中心坐标与螺栓孔中心坐标。第二阶段，根据多个螺栓孔中心坐标与轮毂中心坐标，计算出当前轮毂磨抛点到目标初始点所需的旋转角度。第三阶段，基于模糊控制算法，设计对准控制器，以实现初始点对准控制。通过实验对所提出方法的有效性和精度进行了验证，实验结果表明轮毂磨抛初始点对准的方法是可行的。

关键词: 机器视觉, 轮毂磨抛系统, 模糊控制, 初始点对准