陶瓷零件缺陷的在线视觉检测系统

发布时间：2019-03-10 18:24

【摘要】：陶瓷是生活和生产不可缺少的材料,陶瓷的缺陷会严重影响陶瓷的质量,陶瓷圆环和方形陶瓷管因体积小、批量大给生产过程中的检测带来较大的难度。传统的陶瓷零件检测依靠人工肉眼检测,检测效率低、稳定性差、容易出现误检和漏检现象。近年来,基于声光电等的无损检测方法在陶瓷缺陷检测中引起了高度关注。本文采用机器视觉的办法,针对陶瓷圆环或方形陶瓷管的缺陷检测,探讨一种陶瓷零件缺陷的在线视觉检测方法,主要内容包括:1)在线检测系统的设计,主要包括:图像采集单元、运动控制单元、图像处理单元、人机交互单元等,通过将机械、电机、光学、机器视觉等相关的理论知识和实际应用相结合,设计并实现陶瓷零件缺陷的机器视觉检测系统,可实现对陶瓷圆环和方形陶瓷管缺陷的检测。2)陶瓷圆环检测算法研究,采用投影法+Hough变换的方法对圆形陶瓷进行检测,根据圆形的几何特征,利用投影法对图像进行两次一维扫描就可以得到圆形的圆心坐标,根据Hough变换来对半径进行求解。3)方形陶瓷管检测算法研究,采用投影法+局部分块的方法研究方形陶瓷管的检测方法,利用有效区域来获取图像中的方形陶瓷管,通过方形陶瓷管的重心加权来对其进行倾斜度调整,利用局部分块将方形陶瓷管分成多块,分别判断各块的特征来对其进行缺陷检测。实验结果表明,采用投影法+Hough变换的方法便于检测同心圆,可将Hough变换中参数的三维累计转换为一维累计,减少算法耗时,采用投影法+局部分块的方法简化了方形陶瓷管的检测过程,同时也保证了检测正确率。实验数据表明,陶瓷零件缺陷检测算法耗时83ms,能够达到陶瓷圆环零件的10Hz检测速度要求。
[Abstract]:Ceramics are indispensable materials for life and production. The defects of ceramics will seriously affect the quality of ceramics. Because of the small volume of ceramic torus and square ceramic tubes, large batches of ceramic tubes have brought great difficulty in testing in the process of production. The traditional detection of ceramic parts depends on artificial eye detection. The detection efficiency is low, the stability is poor, and it is easy to appear the phenomenon of false detection and missed detection. In recent years, non-destructive testing methods based on acoustic and optoelectronic have attracted much attention in ceramic defect detection. In this paper, the method of machine vision is adopted to detect the defects of ceramic torus or square ceramic tubes, and an on-line visual inspection method of ceramic parts defects is discussed. The main contents are as follows: 1) the design of on-line inspection system. It mainly includes: image acquisition unit, motion control unit, image processing unit, human-computer interaction unit, etc., through the combination of mechanical, motor, optical, machine vision and other related theoretical knowledge and practical application, The machine vision inspection system of ceramic parts defects is designed and realized, which can detect the defects of ceramic rings and square ceramic tubes. 2) the detection algorithm of ceramic rings is studied, and the method of Hough transformation is used to detect the circular ceramics. According to the geometric characteristics of the circle, the circular center coordinates can be obtained by two one-dimensional scanning of the image by using the projection method, and the radius can be solved according to the Hough transform. 3) the detection algorithm of the square ceramic tube is studied. The detection method of square ceramic tube is studied by using projection method, the effective region is used to obtain the square ceramic tube in the image, and the inclination of square ceramic tube is adjusted by weighting the center of gravity of the square ceramic tube. The square ceramic tube is divided into several blocks by local block, and the characteristics of each block are judged to detect the defects. The experimental results show that the method of Hough transform is convenient to detect concentric circle, and the three-dimensional accumulation of parameters in Hough transform can be converted into one-dimensional accumulation, thus reducing the time-consuming of the algorithm. The projection method is used to simplify the detection process of the square ceramic tube and to ensure the correct detection rate at the same time. The experimental data show that the defect detection algorithm of ceramic parts takes 83 Ms and can meet the requirement of 10Hz detection speed of ceramic ring parts.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TQ174.66;TP391.41

【参考文献】