基于视觉的小型冲压件识别定位研究

摘要

结合机器视觉的优越实时性,对双半圆头小型冲压件的识别与定位进行了研究.相机标定遵循张正友标定法,利用Visual studio2012的开发平台和OpenCV函数库实现.通过视觉系统采集图像,分析对比各类灰度和平滑的预处理方法的效果,确定双边滤波和闭开运算的方法.基于轮廓查找和旋转卡壳算法提取特征参数,基于几何参数的识别算法识别工件类型,由最小面积和最小外接矩得到中心坐标和角度.实验结果表明,小片识别率为100%,外壳、重叠或侧立的工件识别率为97%.该研究识别准确率高,定位精确,自动化和柔性化程度较高,满足装配作业的实际工程要求.%Combining with superior real-time of machine vision,the recognition and location of small semi-circular stamping parts are studied.Camera calibration follows Zhang Zhengyou calibration method,using development platform of Visual studio 2012 and OpenCV function library to achieve.Acquiring images through the visual system,and comparing various types of gray and smooth pre-processing methods to determine the effect of bilateral filtering and open and close operation.Extracting feature parameters based on contour search and rotating block algorithm,identifying types based on geometric parameters ,and the center coordinates and angles are obtained by minimum area and minimum circumscribed torque.The experiment shows that recognition rate of the small slice is 100%,and recognition rate of the shell and overlap or side is 97%.The study identifies a high degree of accuracy,positioning accuracy,high automation and flexibility to meet the practical assembly requirements.