基于机器视觉的纱笼纱杆快速定位方法

为实现纺织印染过程全流程自动化、数字化、智能化生产,针对纺织印染行业大尺寸纱笼上密集纱杆数字化精确定位问题,提出一种离线检测与在线检测相结合的数字化视觉定位方法。该方法将单目工业相机安装在机器人末端,离线检测单元在非生产时间时对纱笼上的纱杆位置进行定位检测,并将纱笼上所有纱杆的位置信息记录到数据库中;在线检测单元在生产过程中对纱笼上少数几个纱杆进行定位检测,结合离线检测的纱杆位置信息,采用最小二乘法计算得到纱笼的整体位姿偏移,然后根据纱笼位姿偏移计算得到所有纱杆的位置信息,从而引导机器人完成相应的取放作业。实验表明,该方法定位准确,鲁棒性强,生产中占用节拍少,实用性强。

Fast location of yarn-bars on yarn-cage based on machine vision

In order to realize automatic, digital and intelligent production in textile dyeing industry, a vision based locating method was proposed for an eye-in-hand system aiming at the problem of location of dense yarn-bars on yarn-cage, combing the offline with online detection information. The offline detection unit recognized and located all the yarn-bars on yarn-cage during the machine idle time, and the location information of the yarn-bars was sent to the database. The online detection unit calculated the positions of all the yarn-bars on the yarn-cage, and guided the robot picking or putting objects on the yarn-bars in the productive time. For the online detection a few yarn-bars were detected firstly, and then a least square method were used to calculate the rotation angle and translation vector of the yarn-cage between its online position and offline position. All the positions of yarn-bars on yarn-cage were calculated using the data from the database. Experiment results show that the proposed method offers accuracy, robustness and practicality.