基于小波熵的织物疵点检测方法

Detection Method of Fabric Defect Based on Wavelet Entropy

探讨利用小波熵检测织物疵点的新方法。在采用二维离散小波变换对织物图像进行分解的基础上,引入了熵的概念,将小波熵作为织物图像的特征值,把熵值看作系统紊乱程度的度量,由此得到织物图像的小波熵特征值,通过与正常织物经过二维小波变换后提取的小波熵值相比较,熵值大者即认为有疵点存在。通过对不同组织织物和几种典型的织物疵点进行检测,试验结果表明该方法有效可行。

New method of using wavelet entropy detecting fabric defect was discussed. The concept of entropy was introduced on the base of using two-dimensional discrete wavelet transform to analyse fabric image. Wavelet entropy was used as eigenvalue of fabric image,entropy was taken as degree of system disorders,then wavelet entropy eigenvalue of fabric image could be got. The larger entropy is considered as defect existing through comparing with wavelet entropy that was got from normal fabric with two-dimensional wavelet transform. The test results show the method is available and effective through detecting different fabric with different fabric texture and several typical fabric defects.