印刷质量的图像检测技术研究

发布时间：2018-05-03 12:29

本文选题：机器视觉 + 图像处理　；参考：《华东理工大学》2017年硕士论文

【摘要】：在制造业转型升级背景下,印刷业面临传统技术改造的难题。质量检测作为印刷生产过程中的必要环节,其检测技术的创新与发展对于传统印刷业变革具有重要的现实意义。为了有效提高印刷质量检测的自动化程度,本文在机器视觉和数字图像处理的基础上,综合运用光源照明、传感器、光学成像、软件开发等知识,研究一种印刷质量的图像检测技术。通过采用模块化设计开发印刷质量检测系统,为解决实际印刷过程中存在的检测效率低、稳定性差等问题提供了可行性方案。基于系统组成和算法设计两个方向,本文详细阐述了印刷质量的图像检测技术在包装印刷领域的应用和系统实现。本课题的主要研究内容如下:(1)针对实际生产需要,通过分析机器视觉原理及其应用,合理确定系统硬件设备选型方案,设计了一套由传送带、LED光源、CCD工业相机、光电开关、旋转编码器和PLC控制器组成的印刷图像采集装置。在采集印刷品图像时,系统利用Pylon Viewer程序驱动相机自动完成对印刷品准确拍摄。针对外界光照对图像采集过程的影响,本装置对相机和光源进行密封操作,从而保证后续图像处理时能够获得高质量的印刷图像。(2)为确保系统顺利实现印刷图像自动检测过程,根据印刷检测的技术要求,本文提出了一系列印刷图像处理识别相关算法,主要包括图像预处理、图像配准和缺陷分类识别三大类。在印刷图像预处理过程中,本文对图像灰度化、图像增强、图像分割等关键算法详细介绍,为后续检测结果的准确性提供保障。根据不同类型印刷品特点,本文提出两种基于ROI模板及基于Hough和Fourier变换的印刷图像配准算法,为进一步缺陷识别奠定良好的基础。针对检测系统的功能需求,本文设计缺陷目标提取与分类识别算法,同时通过改进多类支持向量机完成印刷缺陷的自动识别与分类。最后,本文基于C++编程语言,利用Visual Studio 2013开发工具,综合运用编程和软件项目开发知识,实现印刷质量检测系统可视化平台。
[Abstract]:Under the background of manufacturing industry transformation and upgrading, the printing industry faces the difficult problem of traditional technology transformation. As a necessary link in the process of printing production, the innovation and development of quality inspection technology is of great practical significance to the traditional printing industry. In order to improve the automation of printing quality inspection effectively, this paper, on the basis of machine vision and digital image processing, synthesizes the knowledge of light source lighting, sensor, optical imaging, software development, etc. An image detection technique for printing quality is studied. A printing quality inspection system is developed by modular design, which provides a feasible scheme for solving the problems of low detection efficiency and poor stability in the actual printing process. Based on the two directions of system composition and algorithm design, this paper describes the application and system implementation of printing quality image detection technology in the field of packaging and printing in detail. The main research contents of this subject are as follows: (1) according to the actual production needs, through analyzing the principle and application of machine vision, reasonably determining the selection scheme of the hardware equipment of the system, designing a set of CCD industrial camera and optoelectronic switch with the LED light source of the conveyor belt. The printing image acquisition device composed of rotary encoder and PLC controller. When collecting print image, the system uses Pylon Viewer program to drive the camera to automatically complete the accurate shooting of printed matter. In view of the influence of outside illumination on the image acquisition process, the device seals the camera and light source to ensure that high quality printing image can be obtained in the subsequent image processing. According to the technical requirements of printing detection, this paper proposes a series of printing image processing recognition algorithms, including image preprocessing, image registration and defect classification recognition. In the process of printing image preprocessing, the key algorithms such as grayscale image, image enhancement, image segmentation and so on are introduced in detail in order to guarantee the accuracy of the following detection results. According to the characteristics of different types of printing materials, this paper presents two kinds of printing image registration algorithms based on ROI template and Hough and Fourier transform, which lay a good foundation for further defect recognition. In order to meet the functional requirements of the detection system, this paper designs an algorithm for the extraction and classification of defect targets. At the same time, an improved multi-class support vector machine (SVM) is used to realize the automatic recognition and classification of printing defects. Finally, based on C programming language and Visual Studio 2013 development tools, the visual platform of printing quality inspection system is realized by using the knowledge of programming and software project development.
【学位授予单位】：华东理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TS807;TP391.41

【相似文献】