摘要
在分析织物疵点特点的基础上,提出了一种新的检测方法.首先根据正常织物小波分解后的能量来确定小波基和小波分解层数.接着对测试图像进行小波静态分解,然后提取子图中的子窗口标准差作为提取的特征;在此基础上,通过与正常子图中的特征相比较来判断小疵点是否存在.最后通过提取特征构造的标准差图确定小疵点的位置.实验表明,该检测方法是有效的.
On the basis of analysis of the defect characteristicsof fabrics,a new method for small defect detection was presented.Firstly,the normal fabric energy of wavelet decomposition was used to determine the wavelet func-tion and level of wavelet decomposition.The test image was decomposed statically by wavelet,and then stan-dard deviation of sub-window in the sub-image was extracted as the extracted features,which was compared with the normal sub-image features to determine the existence of defects.Finally,the location of defect was determined by the standard deviation image.Experimental results show that this method is effective.
出处
《天津工业大学学报》
CAS
北大核心
2010年第5期73-76,共4页
Journal of Tiangong University
基金
西安工程大学博士科研基金项目(BS1005)
关键词
小波分解
特征提取
标准差图
疵点检测
wavelet decomposition
feature extraction
standard deviation image
defect detection
