摘要
针对传统方法检测砌墙砖外观缺陷和几何尺寸存在效率低和准确度差的问题,设计一种基于机器视觉的砌墙砖自动检测系统。以OpenCV和PyQt为系统上位机开发平台,首先通过高分辨率的单目工业相机采集砌墙砖图像,再对图像进行阈值分割、边缘提取等图像处理,控制机械臂将标准块置于试样截面形心处,通过检测算法计算出试样的长度、宽度和高度以及识别缺陷区域,其后移开标准块,将缺陷区域坐标转换为运动导轨位移量,控制固定于运动导轨的精密激光传感器扫描缺陷区域实现精确测量缺陷深度数据。该系统已投入检测机构进行试检测,完成多批砌墙砖产品的检测实验,大量的实验结果表明,该检测系统技术成本低、性能稳定、精度高、结果稳定可靠,满足国家标准≤0.5mm的精度要求,尺寸测量重复性为0.09mm,缺陷测试重复性为0.10mm,相比于人工测量分别提高5.4倍和7.8倍,检测时间最多降低99.28%,能快速准确地对砌墙砖完成外观质量和几何尺寸的检测。
For the problem of low efficiency and poor accuracy of traditional methods to detect the appearance defects and geometric dimensions of masonry bricks,a machine vision-based automatic masonry brick inspection system is designed.Using OpenCV and PyQt as the upper computer development platform,the system first acquires masonry brick images by a high-resolution monocular industrial camera,then performs image processing such as threshold segmentation and edge extraction,controls a mechanical arm to place a standard block at the crosssectional center of the specimen,calculates the length,width and height of the specimen and identifies the defect area by the detection algorithm,then removes the standard block and converts the coordinates of the defect area into the displacement of the motion guide.After the standard block is moved away,the coordinates of the defect area are converted into the displacement of the motion guide,and the precision laser sensor fixed to the motion guide is controlled to scan the defect area to achieve accurate measurement of the defect depth data.The system has been put into the testing organization for trial testing,completed a number of batches of masonry wall brick products testing experiments,a large number of experimental results show that the detection system technology,low cost,stable performance,high precision,stable and reliable results,to meet the national standard≤0.5mm accuracy requirements.Compared with manual measurement,the repeatability of dimensional measurement is 0.09mm and the repeatability of defect test is 0.10mm,which is 5.4 times and 7.8 times higher respectively,and the inspection time is reduced by 99.28%at most.
作者
刘付渝杰
宋俊儒
罗睿
许浩然
刘洁利
黎恒平
张昊天
LIUFU Yujie;SONG Junru;LUO Rui;XU Haoran;LIU Jieli;LI Hengping;ZHANG Haotian(Quality&Metrology Supervision Testing Institute,Maoming 525000,China;School of Computer Science and Technology,Guangdong University of Technology,Guangzhou 510000,China;Maoming branch of Guangdong Institute of Special Equipment Inspection and Research,Maoming 525000,China;School of Intelligent Systems Engineering,Sun Yat-Sen University,Shenzhen 518107,China;Pengcheng Laboratory,Shenzhen 518055,China)
出处
《中国测试》
CAS
北大核心
2022年第S02期150-157,共8页
China Measurement & Test
基金
广东省市场监督管理局科技项目(2021CZ30)
关键词
机器视觉
尺寸偏差
外观质量
自动测量
砌墙砖
检验检测
machine vision
dimensional deviation
appearance quality
automatic measurement
masonry bricks
inspection and testing
