论文总字数：21301字

摘 要

近几年来，智能手机行业发展迅猛，人们对于手机屏幕大小、分辨率等显示方面的要求越来越高。小尺寸屏是手机等现代智能终端的关键零部件。在生产过程中手机屏幕表面难免会出现瑕疵，这就对屏幕的质量产生了很大影响。传统的人工肉眼检测方法容易受到环境和个体的主观差异性等因素影响，且稳定性和可靠性难以达到生产企业的触摸屏品质检测和检测效率的要求。

本文通过对传统人工检测观察和对其他屏幕瑕疵检测算法的研究，利用边缘检测和图像分析，检测出小尺寸屏幕瑕疵，通过连通域的特征提取方法分析出瑕疵的特征，从而进行标号，并进行数学形态学的处理，最后通过分析瑕疵的各项特征来判定瑕疵的种类，进行基于机器视觉的方法进行聚类分类识别。利用所给的瑕疵玻璃屏的图像库，进行仿真实验，对比分析各个特征提取与选择算法的有效性。该算法通过了小尺寸屏不良图片库的测试，结果显示界面可以直观的看到测试结果，分析后得出该系统具有较高的准确度和实用性。在此基础之上，该系统可以基本取代传统的人工肉眼检测，能够有效的在生产中控制了小尺寸屏的质量，统一品质标准，同时避免了客户和生产中后期的问题产生，节约资源和成本，提高了企业的核心竞争力，有利于小尺寸屏幕产业的进一步发展。

关键词：机器视觉 小尺寸屏 边缘检测 瑕疵检测 特征提取

Design and implementation of visual inspection algorithm

for small size screen defects

Abstract

In recent years, the smart phone industry has developed rapidly, and the demand for mobile phone screen size, resolution and other display has become higher and higher. Small size screen is a key component of modern smart terminals, such as mobile phones. In the production process, the surface of mobile phone screen will inevitably appear flaws, which will have a great impact on the quality of the screen. The traditional method of artificial eye inspection is easy to be affected by the environment and individual subjective differences, and the stability and reliability are difficult to meet the requirements of the quality inspection and detection efficiency of touch screen.

Based on the traditional artificial detection and to observe the other screen defect detection algorithm, using the edge detection and image analysis, detect the small screen size defects, defect analysis method feature by feature connected domain extraction, and labeling, processing and mathematical morphology, finally through the analysis of the characteristics of defects to determine the types of defects, clustering classification and recognition method based on machine vision. Using the image library to screen glass defects, simulation experiment, comparative analysis of various feature extraction and selection method is effective. The algorithm through the small size screen bad picture library test results show field.You can see the visual surface test results, analyzed the accuracy and practicability of the system is higher. On this basis, the system can replace the traditional artificial eye detection, can effectively control the quality in the production of small size screen, unified quality standards, at the same time to avoid late customers and production. The problem of resource saving and cost, improve the core competitiveness of enterprises, is conducive to the further development of the small size of the screen industry.

Key words:Machine vision;Small size screen;Edge detection;Defect detection;Feature extraction

目 录

摘要 I

ABSTRACT II

第一章 绪论 1

1.1前言 1

1.1.1小尺寸屏缺陷的视觉检测的研究背景 1

1.1.2小尺寸屏缺陷检测发展 1

1.2小尺寸屏缺陷的视觉检测的研究现状 2

1.2.1国外研究现状 2

1.2.2国内研究现状 3

1.3课题研究的内容、手段以及意义 3

1.3.1课题研究的内容 3

1.3.2课题采用的研究手段 4

1.3.3课题的研究意义 4

1.4论文的章节安排 5

第二章 技术基础 6

2.1机器视觉 6

2.2 数字图像处理 6

2.2.1图像边缘检测 6

2.2.1.1 Sobel算法 7

2.2.1.2 Laplace算法 8

2.2.1.3 Canny边缘检测算法 8

2.2.2 K-means聚类算法 10

2.3本章小结 12

第三章 小尺寸屏缺陷检测系统与算法 13

3.1小尺寸屏缺陷的视觉检测的系统框架 13

3.1.1图像采集模块 13

3.1.2相机标定模块 13

3.1.3缺陷检测模块 13

3.2小尺寸屏缺陷检测算法设计 14

3.2.1图像预处理 14

3.2.1.1均值滤波 14

3.2.1.2边缘检测 15

3.2.2瑕疵检测 17

3.2.2.1图像分割 17

3.2.2.2瑕疵标记 18

3.2.3瑕疵分类 19

3.2.3.1面积分类 19

3.2.3.2气泡识别 20

3.3本章小结 21

第四章 小尺寸屏缺陷检测系统仿真与分析 22

4.1仿真结果 22

4.1.1狭小点状瑕疵 22

4.1.2斑状瑕疵 23

4.1.3大面积瑕疵 24

4.1.4气泡 25

4.2结果分析 26

4.3本章小结 26

第五章 总结与展望 27

5.1论文总结 27

5.2未来展望 27

参考文献 29

致谢 31

第一章 绪论

1.1前言

1.1.1小尺寸屏缺陷的视觉检测的研究背景

请支付后下载全文，论文总字数：21301字