基于计算机视觉的织物疵点检测技术研究进展

摘要

对基于计算机视觉的织物疵点检测技术进行回顾,介绍了灰度共生矩阵法,局部二值模式算法,邻域关联分析,自组织映射,支持向量机,学习向量量化分类器,多分类器组合和决策融合等算法等在图像预处理,特征提取、分类和识别等方面的应用情况,着重讨论了一种基于多数投票原则的多分类器决策融合技术,试验结果证实该技术有较高精确性.%In this paper, various methods of fabric defect dectection based on computer vision are reviewed.This article describes the image preprocessing, feature extraction, classificaition and identification for the various algorithms, such as Gray level co-occurrence Matrix, Local Binary Patterns, Contextual Analysis,Self-organizing Maps, Support Vector Method, Learning Vector Quantization, Multi-classifier combination,decision fusion, etc. A multi-classifier decision fusion technique based on majority voting is mainly discussed.Empirical results indicate the high accuracy of the presented approach.