基于机器视觉的布匹疵点检测系统  

Fabric Defect Detection System Base on Machine Vision

作　　者：毕明德[1] 孙志刚[1] 李叶松[1] 

机构地区：[1]华中科技大学控制科学与工程系,图像信息处理与智能控制教育部重点实验室,湖北武汉430074

出　　处：《仪表技术与传感器》2012年第12期37-39,125,共4页Instrument Technique and Sensor

基　　金：国家自然科学基金资助项目(61074124);图像信息处理与智能控制教育部重点实验室开放基金(200702)

摘　　要：为了提高布匹疵点检测的精度与速度,提出了一种基于机器视觉的布匹疵点检测系统用于取代人工检测。论述了系统的整体结构,包括成像设备、光源选择以及图像采集与处理方式等,并提出了一种基于类别共生矩阵与支持向量机的布匹疵点检测算法。检测算法将疵点检测看作一个两类分类问题,采用从灰度共生矩阵中提取的特征来描述纹理特性,并采用支持向量机来对特征向量进行分类完成疵点的检测。最后通过大量的布匹疵点实例对算法的可靠性进行验证,并对检测算法在不同参数下的检测精度与实时性进行了讨论。To improve the detection accuracy and speed of the fabric defect detection, a machine vision based detection system was proposed to instead the human inspector. The architecture of the system was elaborated including the imaging device, illumina- tion method, and image grabbing and processing methods. A new fabric defect detection algorithm was also proposed based on gray- level co-occurrence matrix and support vector machine. Fabric defect detection was considered as a two-class classification prob- lem. The features extracted from gray-level co-occurrence matrix were used to characterize the fabric texture, and the support vector was used as a classifier for the extracted features to decide whether the fabric was defective or not. The detection results of large quantities of images acquired on industrial occasions were provided to validate the usefulness of the proposed algorithm. And the de- tection accuracy and the real-time performance of the algorithm with different parameters were also discussed.

关 键 词：机器视觉 疵点检测 灰度共生矩阵 支持向量机 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]