摘要
织物疵点部分相当于图像中灰度突变的奇异信号,对比传统的边缘检测方法,小波变换在检测图像微小细节边缘时具有明显的优势,能很好地刻划突变点的奇异性。首先利用小波的多尺度特性求得织物图像局部极大值点,有效提取出织物疵点区域的边缘,然后结合形状特征求出织物疵点区域的形状特征,对常见重纬、重经、缺经、缺纬、油污、破洞织物疵点图像进行仿真实验,结果表明此方法既保留了织物疵点边缘信息,又剔除了虚假边缘,最终有效地提取出了疵点区域的形状特征。
Fabric defect is singularity signal of the gray image,compared with the traditional test methods,wavelet transform has obvious advantages in detection of minute edge details of image,and able to describe singularity better.first,use multiscale characteristic of wavelet transform to find local modulus maxima of fabic image,effectively obtain the edge of fabric defect parts,then combine morphology to find appearances of the fabric defect,experiment on common defect images of coarse pick,coarse end,end out,mispick,clip mark and oil blemish,experiment results show that this method can retain defects edge information effectively and can remove false edge,eventually,effectively obtain the shape characteristics of the fabric defect parts.
出处
《电子测量技术》
2010年第10期108-110,共3页
Electronic Measurement Technology
关键词
小波变换
模极大值
形态学
织物疵点
wavelet transform
maxima modulus
morphology
fabric defect
