基于AR谱分析的织物瑕疵自动检测  

Fabric Defect Detection Based on AR Spectral Estimation and Support Vector Data Description

作　　者：步红刚[1] 汪军[1,2] 黄秀宝[1,3] 

机构地区：[1]东华大学纺织学院,上海201620 [2]东华大学纺织面料技术教育部重点实验室,上海201620 [3]东华大学纤维材料改性国家重点实验室,上海201620

出　　处：《数据采集与处理》2009年第6期783-788,共6页Journal of Data Acquisition and Processing

摘　　要：为实现快速和有效的织物瑕疵自动检测,提出了一种基于时间序列而不是图像的功率谱纹理分析方法。依据Burg au to regress ive(AR)算法估计得到谱数据,从中提取能够反映纹理周期和取向等特点的特征,并首次采用支持向量数据描述模型来检测织物瑕疵纹理。对包含多种疵点的若干织物样本的检测结果表明,依照本文所述方案能够在保持较低的误警率前提下达到较高的疵点检出率,证明了所述方案的可行性。For realizing the fast and effective detection of fabric defects, this paper proposes a novel texture analysis approach based on the power spectral density of one-dimensional time series rather than on the two-dimensional image. By using Burg auto-regressive（AR） algorithm to estimate the spectral density, feature characterizing periodicity and the orientation of the fabric texture are extracted. The support vector data description model is firstly used to detect fabric defect textural. Experimental results indicate that the low false alarm rate and the low missing rate can be simultaneously obtained, thus providing the effectiveness of the proposed approach.

关 键 词：纹理分析 织物瑕疵检测 时间序列分析 AR谱分析 支持向量数据描述 

分 类 号：TS101.9[轻工技术与工程—纺织工程]