中文摘要：

针对传统对位装置对位精度低、速度慢的缺点，设计并开发了基于机器视觉的柔性印刷电路（FPC）智能对位装置。装置利用高分辨率工业CCD相机采集图像，配合机械和电气系统，控制运动控制卡完成FPC智能抓取、采集、处理和对位。在关键的定位算法中，通过改进的一维Otsu阈值分割方法对FPC图像进行阈值分割，利用提出的三椭圆平均拟合法提取定位圆圆心坐标，采用累计概率Hough直线变换测量旋转的角度；进而将计算的相对于标准位置的偏移量传送至运动控制卡，利用闭环反馈控制伺服系统进行纠偏。在开发的样机上进行大量的在线实验结果表明，本文装置稳定性强，对位精度高，对位误差不超过±0．06mml，角度误差不超过0．0016rad。

英文摘要：

Aiming at the disadvantages of low precision and slow speed of the traditional alignment de- vice,a flexible printed circuit （FPC） intelligent alignment device is designed and developed based on ma- chine vision in this paper. Device uses high resolution industrial CCD camera to capture images, and help- ing with the mechanical and electrical parts,it can control the motion control card to complete FPC intel- ligent grabbing,acquisition,processing and positioning. In the key location algorithm,the threshold seg- mentation of FPC image is mainly achieved using the improved OTSU algorithm, the average fitting of three ellipses is used to obtain the coordinates of center of location circle,and the rotation angle is meas- ured using cumulative probability hough transform. The calculated offset relative to the standard position is transmitted to the motion control card, and the closed-loop feedback control system is used for correc- ting the deviation. Finally,a large number of online experiments were carried out on the prototype. The experimental results show that the device has strong stability and high alignment accuracy, the alignment error is not more than 0. 06 ram,and the angle error is not more than 0. 0016 rad.