基于机器视觉的柴油机针阀体智能化加工单元设计与研究  

Design and Research of Intelligent Processing Unit for Diesel Engine Needle Valve Body Based on Machine Vision

作　　者：文传奇 罗付强 何友苗 刘宏岩 WEN Chuanqi;LUO Fuqiang;HE Youmiao;LIU Hongyan(Chongqing Hongjiang Machinery Co.,Ltd.,Chongqing 402160,China)

机构地区：[1]重庆红江机械有限责任公司,重庆402160

出　　处：《内燃机》2023年第3期18-24,共7页Internal Combustion Engines

摘　　要：针对针阀体喷孔加工对零件尺寸在线测量与装夹方向识别的需求,设计了一套基于OpenCV的机器视觉检测的智能化加工单元方案,通过机器视觉对零件进行相关尺寸的测量及装夹方向识别,实现针阀体喷孔加工的全自动化。机器视觉检测的主要流程为:首先,对图像进行畸变处理和灰度化、ROI提取、滤波等图像预处理;再通过Canny算子获得各圆形轮廓像素级的圆心和半径;最后,基于Zernike矩求得亚像素级的轮廓边缘点,以像素级圆心和半径作为约束条件,采用最小二乘法拟合得到亚像素级圆心和半径。经应用表明,智能化加工单元满足使用需求。In response to the demand for online measurement of part size and recognition of clamping direction in needle valve body spray hole processing,an intelligent machining unit scheme based on OpenCV machine vision detection was designed.Through machine vision,the relevant dimensions of the parts were measured and clamping direction recognition was achieved,achieving full automation in needle valve body spray hole processing.The main process of machine vision detection is as follows:first,image distortion processing,graying,ROI extraction,filtering,and other image preprocessing are performed on the image;Then,the center and radius of each circular contour pixel level are obtained through the Canny operator;Finally,based on Zernike moments,sub pixel level contour edge points are obtained,and the pixel level center and radius are used as constraints.The least squares method is used to fit the sub pixel level center and radius.The application shows that the intelligent processing unit meets the usage requirements.

关 键 词：机器视觉 在线检测 图像处理 边缘检测 加工单元 

分 类 号：TK422[动力工程及工程热物理—动力机械及工程]