基于OpenCV的提花机零件识别与定位方法研究
Research on the Identification and Positioning Method of Jacquard Machine Parts Based on OpenCV

本文根据某公司制造的提花机零部件，在智能装配上料过程出现了翻转、遮挡、偏斜和复杂背景等问题，对基于OpenCV双目立体视觉零件的辨识和定位展开了深入研究。在本文中建立了双眼视觉系统模型，然后完成了对双目视觉控制系统的校准，并获得了照相机的内部参量、相应位置和失真比率等参数。然后在ORB特性提取计算结果的基础上，又提出了ORB + RANSAC特性匹配计算，通过选择正误匹配，并通过立体匹配获得视差图，并根据校准计算结果对工具实现了三维重构。实验结果表明本文使用的改进ORB + RANSC属性匹配计算，可以更精确、高效地对工件加以辨识和定位。
In this paper, according to the parts of jacquard machine manufactured by a company, there are problems such as turning, blocking, skew and complex background in the intelligent assembly loading process, and the identification and positioning of the parts based on binocular stereo vision based on OpenCV are deeply studied. In this paper, the binocular vision system model is established, and then the binocular vision control system is calibrated, and the internal parameters, corresponding positions and distortion ratio of the camera are obtained. Then, on the basis of the ORB feature extraction calculation results, the ORB + RANSAC feature matching calculation was proposed, and the parallax map was obtained by selecting the right and wrong matching and stereo matching, and three-dimensional reconstruction of the tool was realized according to the calibration results. The experimental results show that the improved ORB + RANSC attribute matching calculation used in this paper can identify and locate the workpiece more accurately and efficiently.