基于偏微分方程的带钢表面图像去噪方法  

Image denoising method on surface of steel strip based on partial differential equations

作　　者：李长乐[1] 臧希吉吉 程万胜[1] 赵杰[1] 

机构地区：[1]哈尔滨工业大学机器人研究所,哈尔滨150080

出　　处：《大连海事大学学报》2008年第1期71-74,共4页Journal of Dalian Maritime University

基　　金：长江学者和创新团队发展计划资助项目(IRT0423)

摘　　要：为提高带钢表面缺陷图像去噪处理中滤波器的去噪性能,更好地保留缺陷图像细节信息,在研究了偏微分方程理论的基础上,分析了AMSS(affine morphological scale space)方程的对比不变性和仿射不变性,采用有限差分法实现了AMSS算子的方法.将AMSS算子应用到带钢表面缺陷图像去噪处理的试验表明,该方法达到了既去除噪声,又保留图像细节信息的效果.Affine morphological scale space （AMSS） arithmetic operators were obtained by adopting finite difference method to improve denoising performance of the filter in surface defect image of steel strip and maintaining the detail information of the defect image. The classical invariant property and affine invariant property were analyzed based on theory of partial differential equation. AMSS arithn：etic operators were tested to eliminate noise in surface defect image of steel strip, and results show that the method can eliminate noise while maintaining detail informations.

关 键 词：带钢 表面缺陷 图像去噪 偏微分方程 AMSS方程 

分 类 号：TN911.7[电子电信—通信与信息系统]