中文摘要：

针对传统单向器星轮人工装配效率低下问题,分析了单向器星轮装配过程,构建了基于计算机视觉技术的工业机器人自动化装配方案;首先利用CCD摄像机捕捉星轮表面形貌图像,然后利用Open CV视觉库进行了摄像机标定,再结合图像噪声去除,Hough圆检测,轮廓提取等技术,实现了星轮位姿的非接触式视觉检测;最后以单向器为例验证了该检测方法的有效性,机器人通过视觉获取的星轮位姿信息完成自动化装配.

英文摘要：

This paper designs an industrial robot automatic assembly system which solves the problem of traditional star wheel manual assembly inefficiencies, after analyzing the assembly processing of the star wheel. Firstly, the star wheel surface profile image was captured by using CCD camera. Then, camera calibration was achieved with the help of Open CV visual library. Afterwards, the Noncontact visual in- spection is implemented with the help of vision technology such as noise removal, Hough circle detection, contour extraction. At last, in the case of a star wheel, the experiment verifies the effectiveness of the detection method and the industrial robot realizes the automatic assembly processing based on the information obtained by the visual detection system.