基于频域分析的毛巾织物疵点检测

Defect Detection of Towel Fabric Based on Frequency Domain Analysis

摘要:

本文针对毛巾织物的疵点检测进行研究。通过研究分析发现,大部分毛巾织物的疵点主要集中在毛巾表面的毛圈特征上。本文根据疵点和毛圈轮廓特征的频谱特性进行疵点检测,提出了基于频域分析的毛巾织物图像疵点检测算法。以毛巾织物图像频谱的能量统计特性作为疵点检测的特征向量,采用了支持向量机(SVM)作为疵点分类识别算法。matlab仿真实验表明,可以达到平均95%的毛巾织物疵点识别率,说明该频域能量特征提取方法在对纯色毛巾织物的疵点检测上简单有效。

Abstract:

The defect detection of towel fabrics was studied in this paper. Through research and analysis, it was found that most of the defects of towel fabrics are mainly concentrated on the characteristics of the towel surface. According to the relationship between the defect and the contour feature of the towel, the defect detection algorithm of towel fabric based on frequency domain analysis was proposed. The energy statistical characteristics of towel fabric image spectrum were used as feature vectors for defect detection. Support Vector Machine (SVM) was used as defect classification and recognition algorithm. Matlab simulation results showed that the average recognition rate of towel fabric defects can reach 95%. The conclusion showed that the energy feature extraction method in frequency domain is simple and effective in defect detection of pure color towel fabric.