拟合在布匹瑕疵检测技术中的应用

发布时间：2018-03-31 23:38

本文选题：布匹瑕疵检测　切入点：LBP算法　出处：《吉林大学》2015年硕士论文

【摘要】：随着生活节奏的加快，以及生活质量的提高，我们对衣食住行的方方面面都有了更高的要求．衣服除了有帮助我们抵御寒冷、保护身体等作用以外，在现代社会它还起到了对人体的装饰作用，象征了一个人的生活品味、社会地位和消费水平等等．人们对衣物美观上的追求，不得不使人们注重布匹质量的好坏．此外，纺织工业作为我国的支柱型产业，它的发展对国民经济起着至关重要的作用．对于纺织工业而言，高质量的纺织品是企业盈利的重要因素之一．因此，在生产过程中，需要对布匹的质量进行严格的把控．由于人们的生活需求和纺织工业的发展需求，布匹瑕疵检测技术越来越受到人们的重视．随着计算机技术的飞速发展，布匹瑕疵的检测技术也逐步从人工检测过渡到计算机智能检测．关于布匹瑕疵检测这一课题的研究背景，本文第一章给予了详细的介绍．随后介绍了几种瑕疵检测与分类算法在布匹检测中的应用，以及国内外关于布匹检测技术的研究进展．其中算法包括基于局部二进制模式()的改进算法，以及它的实际应用和实验结果．结合国内某工厂的的实际情况，本文第二章提出了一种新的方法来解决实际布匹生产中的检测问题．将曲线拟合应用到瑕疵检测中，来避免非均匀光照对于布匹瑕疵检测的影响，使得瑕疵部分更突出，更方便阈值的选取，从而达到更有效更快捷的检测．最小二乘法作为一种数学优化技术，将其应用于灰度图像的拟合，实验证明，对于几种特定的瑕疵，能达到很好的检测效果．本文对实际问题进行了具体分析并且做了大量的实验．为了达到更好的拟合效果，对待检测图像进行了适当的分割，，选取了最合适的多项式阶数，并结合K-means算法确定了恰当的阈值．本文完成了布匹瑕疵点的在线检测．
[Abstract]:With the quickening pace of life and the improvement of the quality of life, we have higher requirements on all aspects of clothing, food, housing and transportation. In modern society, it also plays a decorative role to the human body, symbolizing a person's taste in life, social status and consumption level, etc. People's pursuit of clothing aesthetics has to make people pay attention to the quality of cloth. As the pillar industry of our country, the development of textile industry plays a vital role in the national economy. For the textile industry, high quality textiles is one of the important factors for enterprises to make profits. Therefore, in the process of production, It is necessary to strictly control the quality of cloth. Because of the needs of people and the development of textile industry, people pay more and more attention to the technology of fabric defect detection. With the rapid development of computer technology, Fabric defect detection technology is gradually from manual detection to computer intelligent detection. The first chapter gives a detailed introduction to the research background of fabric defect detection, and then introduces the application of several defect detection and classification algorithms in fabric detection. And the research progress of cloth detection technology at home and abroad, including the improved algorithm based on local binary mode), its practical application and experimental results. According to the actual situation of a factory in China, a new method to solve the problem of testing in the production of fabric is put forward in the second chapter of this paper, and the curve fitting is applied to the detection of defects. To avoid the influence of non-uniform illumination on fabric defect detection, make the defect part more prominent, more convenient to select the threshold, so as to achieve more effective and faster detection. The least square method as a mathematical optimization technology, It is proved by the experiment that it can achieve good detection effect for several special defects. In this paper, the practical problems are analyzed and a lot of experiments are done. In order to achieve better fitting effect, the method is applied to the fitting of gray-scale image. The detection image is segmented properly, the most appropriate polynomial order is selected, and the appropriate threshold is determined with K-means algorithm. In this paper, the on-line detection of fabric defect points is completed.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.41

【共引文献】