CAJ | 학술논문

提出一种由灰度共生矩阵生成相应特征图像的算法，进行了图像分割和织物疵点检测。先将织物疵点图像进行灰度级量化至16级，再提取0°，45°，90°，135°四个方向上的灰度共生矩阵，通过计算灰度共生矩阵中的熵、相关性、对比度、差异性、逆差矩共五种特征值并生成相应的特征图像，对常见的5种织物疵点进行了分割检测实验。实验结果证明基于灰度共生矩阵生成特征图像的检测算法是一种检测效果良好的疵点检测方法。
제출일충유회도공생구진생성상응특정도상적산법，진행료도상분할화직물자점검측。선장직물자점도상진행회도급양화지16급，재제취0°，45°，90°，135°사개방향상적회도공생구진，통과계산회도공생구진중적적、상관성、대비도、차이성、역차구공오충특정치병생성상응적특정도상，대상견적5충직물자점진행료분할검측실험。실험결과증명기우회도공생구진생성특정도상적검측산법시일충검측효과량호적자점검측방법。
The fabric defect detection algorithm that feature images were generated by gray level co-occurrence matrix feature image was proposed. This paper quantized the gray-level of the fabric defect image to 16 , then extracted the gray level co-occurrence matrix in these 0° , 45° , 90° , 135° four direction. Through calculating five kinds of characteristic value including entropy, correlation, contrast, differences and inverse matrix of gray level co-occurrence matrix and generaing the corresponding feature image, to make five common types of fabric defects are segmentaed and detected. Experiment results prove that the detection algorithm based on gray level co-occurrence matrix is a good fabric defect detection method.