Aiming at the problem of insufficient positioning accuracy of the ring spinning spun yarn automatic jointing machine using an automated guided vehicle platform after spun yarn breakage, a three-dimensional coordinate detection method of the ring spinning spun yarn spindle position based on binocular vision is proposed. The bottom column of the spindle was selected as the detection target, the contrast of the background of the bottom column of the spindle was physically enhanced, the picture of the bottom of the spindle was collected with a calibrated binocular camera, the two-dimensional coordinates of the midpoint of the bottom edge of the spindle bottom profile were obtained by the processing of the local picture, and the three-dimensional world coordinates of the midpoint of the bottom edge of the spindle bottom profile were calculated based on the method of single-pair feature-point matching. Comparing this paper’s single-pair feature-point matching method with the traditional multi-pair feature-point matching detection method (semi-global block matching), the experimental results showed that: the detection errors of both the single-pair feature-point matching method and the traditional semi-global block matching detection method were less than 2 mm; both methods satisfied the requirements of the positional detection accuracy of the spindle joints. The position detection method based on single-pair feature-point matching saved 80% of the program time compared with the traditional semi-global block matching method.

中文翻译：

---

基于双目视觉的环锭纺细纱锭子位置测量

针对采用自动导引车平台的环锭纺细纱自动接头机细纱断头后定位精度不够的问题，提出了一种基于双目视觉的环锭纺细纱锭子位置三维坐标检测方法。选择主轴底柱作为检测目标，对主轴底柱背景进行物理增强对比度，用标定后的双目相机采集主轴底部图像，得到二维坐标通过对局部图片的处理得到主轴底部轮廓底边中点的三维世界坐标，并基于单坐标的方法计算出主轴底部轮廓底边中点的三维世界坐标。对特征点匹配。将本文的单对特征点匹配方法与传统的多对特征点匹配检测方法（半全局块匹配）进行比较，实验结果表明：单对特征点匹配的检测误差方法与传统半全局块匹配检测方法的误差均小于2mm；两种方法均满足主轴关节位置检测精度的要求。基于单对特征点匹配的位置检测方法与传统的半全局块匹配方法相比节省了80%的程序时间。