摘要
采用了二维离散小波变换对织物图像进行处理,根据小波变换后高、低频分量的特点,提取分解子图像的特征值,送入BP神经网络进行识别,从而实现织物疵点的检测。
Fabric images are dealt with two-dimensional wavelet transform. The characteristics of the decomposed sub images are drawn, the values of which are trained and recognized by BP neural network so that the fabric defect can be detected.
出处
《浙江工程学院学报》
2004年第1期15-19,共5页
Journal of Zhejiang Institute of Science and Technology
