摘要
提出一种利用多层小波和共生矩阵进行纹理表面缺损检测的有效方法.该方法首先将缺损图像在不同水平上进行小波分解,得到一系列低频子图像和高频细节子图像;然后计算和分析各水平上高频细节子图像的共生矩阵特征;最后选择低频子图像进行小波合成得到无纹理图像进行检测.实验证明,该方法能够快速准确地进行纹理缺损检测.
An efficient approach of using multi-level wavelet transform and co-occurrence matrix for texture defect detection was proposed. The defective image is firstly decomposed into approximation and detail sub-images at various levels by wavelet transform. Then, the co-occurrence matrix features of the detail sub-images are computed and analyzed to decide the appropriate level at which the approximation sub-images are reconstructed into non-texture images for defect detection. The experimental results show that this approach is efficient both in speed and performance.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2006年第3期425-430,共6页
Journal of Shanghai Jiaotong University
基金
上海市科委计划项目
关键词
图像识别
小波变换
共生矩阵
纹理分类
缺损检测
image recognition
wavelet transform
co-occurrence matrix
texture classification
defect detection