【作者】 刘宁；

【导师】 秦付军； 胡国高；

【作者基本信息】 西华大学 ， 机械工程， 2016， 硕士

【摘要】 随着科学技术的发展以及社会的进步,越来越多的产品制造厂商更加注重品牌logo的设计与制造,从而提高企业的品牌形象和增强产品的软实力。为了确保自身产品logo的外观品质,对其表面质量的缺陷检测成为一项生产厂家十分关注的课题。本文以某厂家采用铝合金板材加工而成的产品logo为检测对象,根据客户提出的表面质量检测要求,提出了一种基于机器视觉的logo表面缺陷检测方案。对待检测logo表面光滑、缺陷微小等特性进行分析后,搭建了一套由韩国IMI Technology公司生产的高分辨率CCD千兆网口相机、SHI-PCIE-541T2双口千兆网卡、日本VS公司生产的VS-LZ0510/F工业定焦镜头、LED圆顶型无影光源以及计算机组成的机器视觉缺陷检测硬件平台。在此平台的基础上采集待检测logo的图像,并通过数字图像算法处理后判断其表面是否存在缺陷及缺陷的类别。本系统利用LabVIEW作为图像处理算法的软件平台,LabVIEW有着虚拟仪器特殊的优势,充分利用了其图形化编程语言的高效性与便捷度,通过NI Vision视觉函数模块与MATLAB混合编程的方法,实现了对图像的采集、预处理、阈值分割、形态学处理、掩膜、缺陷识别等功能,有效的抑制了背景干扰,提取缺陷特征,得到待检测logo的表面是否存在缺陷的判断,并根据提取到缺陷图像的特性予以分类。最后将处理结果通过串口通信方式传输给可编程控制器S7-200PLC,PLC可根据接收到的信号做出相应的运动控制。经试验检测,本系统能够准确、高效地得到待测logo是否合格的判断,并根据缺陷特征对logo的表面瑕疵进行分类。基于机器视觉的表面缺陷检测方式,相比传统的人工目测检测方式大大提高了检测精度,确保了检测标准的一致性。基于机器视觉的logo表面缺陷检测,是一种高效率、高精度的检测方式,提高生产效率的同时降低了劳动成本。更多还原

【Abstract】 With the development of science and technology and the progress of the society,more and more manufactures pay more attention to the logo’s design,so as to improve the impact of their products and enhance the soft power of the company.In order to ensure the logo’s appearance quality,manufacturers pay more attention to the defect inspection of surface quality.In this thesis,the object to be detected is a product logo machining from aluminum alloy sheet,according to the surface quality detection scheme based on Machine Vision was built,which composed of a high resolution GigE CCD produced by IMI Technology,Korea;SHI-PCIE-541T2 dual-port NIC;VS-LZ0510/F Industrial prime lens produced by VS,Japan;LED Dome-type shadowless light source and data processing computer.Get logo images with this image acquisition device,and through the digital image processing algorithms to determine whether the surface defects and defect categories.The system uses LabVIEW as a software platform for image processing algorithms,it has high efficiency and convenience of the graphical programming language.Through NI Vision function module and MATLAB mixed programming method,this object realized image acquisition,preprocessing,thresholding,morphological processing,mask defect identification and other functions.The system can effectively suppress the background noise thereby extracting defect characteristics,determining whether the logo’s surface have defects or not and classifying the defect based on its characteristic parameters.Finally,the results would be transmitted to the programmable controller S7-200 PLC through serial communication,then PLC could make the appropriate motion control based on the received signal.Experimental results show that the system can accurately and efficiently determine logo surface quality qualified or not,and according to the characteristics of the logo of the defect classifying surface defects.Compared with traditional manual inspection,the method based on Machine Vision greatly improves the detection accuracy of the testing standards to ensure consistency.This object based on Machine Vision for logo’s surface defect,it is a high-efficiency,high-precision testing method,improving productivity while reducing labor costs.更多还原