【作者】 王清；

【导师】 张树功；

【作者基本信息】 吉林大学 ， 计算数学， 2015， 硕士

【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．更多还原

【Abstract】 With the speeding up of the pace of life as well as the improvement of the qual-ity of life,we have a higher request of all aspects of food,clothing,shelter and meansof traveling.Clothes help us to cope with the cold and protect the body,other thanthat,in the modern society,clothes also have played an important role to the humanbody’s decoration and the symbol of a people’s life taste,social status,consumptionlevel and so on.Because of people’s pursuit in the artistic clothes,that people haveto pay attention to the question about fabric quality.In addition,textile industry asa pillar industry of our country,its development has played a vital role to the na-tional economy.To the textile industry,high quality textile is one of the importantfactors of the corporate profits.Therefore,in the process of production,people needto control the fabric quality strictly.On account of the demand of people’s life andthe development of textile industry,fabric defect detection more and more gets peo-ple’s attention.With the rapid development of computer technology,the technologyof fabric defect detection transforms from artificial detection to computer intelligentdetection step by step.The research background of the task about fabric defect detection,this papermakes a detailed introduction in the first chapter.Then I will introduce the applica-tion of several defect detection and classification algorithms and the survey aboutthe technology of fabric defect detection at home and abroad.In this chapter,the in-troduction of algorithms include the improved LBP algorithm based on local binarypattern and its practical application and experimental results.Combined with the actual situation of a domestic factory,in the second chap-ter,a new method to solve the detecting problem in the actual production of fab-ric is proposed.Applying curve fitting to the defect detection to avoid the impactof non-uniform illumination to fabric defect detection,to make the defective partsmore outstanding and the selection of threshold more convenient,so as to achievemore efcient and fast detection.The least square method as a kind of mathemat- ical optimization techniques,is applied to the fitting of gray image,the experimentproved that it can achieve very good detection efect for several special defects.Thispaper makes a concrete analysis to the actual problem and does a large numberof experiments.In order to make better fitting efect,I make a segmentation to thetreating image correctly and select the most appropriate number of polynomial anddetermine an optimum threshold combined with K-means algorithm.This paper hascompleted the fabric defect detection online.更多还原