摘要
在中值滤波的基础上,提出了小波包和图像融合的织物疵点图像预处理方法。利用小波包进行图像全局阈值降噪,分别采用软、硬的阈值方法得到两类图像,并将两类图像进行小波分解融合。结果表明,该方法可以很好地消除噪声和不必要的信息,同时保留了疵点信息。
On the basis of median filtering, this thesis proposes a preprocessing method of fabric defect image using wavelet packet and image fusion. Then, it selects global threshold in wavelet packet de-noise. Two kindsof images are acquired by soft and hard threshold, respectively. The two kinds images are decomposed and fused by wavelet. The results show that the method can eliminate the noise and unnecessary information as well as retain the defect information.
出处
《丝绸》
CAS
北大核心
2012年第8期30-33,共4页
Journal of Silk
基金
江苏省高校优势学科一期建设工程项目(苏政办发[2011]6号)
现代丝绸国家工程实验室开放课题(2010007)
关键词
织物疵点
小波包
图像融合
图像预处理
Fabric defect
Wavelet packet
Image fusion
Image preprocessing
