本文二维码信息
二维码（扫一下试试看！）
基于多尺度冗余字典稀疏分解的纸病图像背景补偿方法
Background Compensation Method of Paper Defects Image Base on Sparse Decomposition via Multi-scale Redundant Dictionary
收稿日期：  
DOI：10.11980/j.issn.0254-508X.2016.11.010
关键词:  纸病检测  稀疏分解  正交匹配追踪算法  多尺度冗余字典  图像背景补偿
Key Words:paper defects detection  sparse decomposition  OMP  multi-scale redundant dictionary  image background compensation
基金项目:陕西省科技攻关项目(2016GY-005);陕西省科技统筹创新工程计划项目(2012KTCQ01-19);陕西省科技攻关项目(2011K06-06);西安市未央区科技计划项目201304。
作者单位
周　强 陕西科技大学电气与信息工程学院,陕西西安,710021 
杜晞盟* 陕西科技大学电气与信息工程学院,陕西西安,710021 
王志强 陕西科技大学电气与信息工程学院,陕西西安,710021 
摘要点击次数: 5691
全文下载次数: 1904
摘要:对于纸病检测中纸张图像背景不均匀以及图像灰度特征不明显等造成纸病测量精度低的问题，建立多尺度冗余字典,采用正交匹配追踪算法(OMP)对纸病图像进行稀疏分解,并根据纸病背景图像和纸病图像不同形态特征,对背景进行补偿,从而增强纸病特征。实验表明,该方法能够有效地重构并补偿纸病背景图像,突出灰度特征较弱的纸病,提高纸病检测的准确性。
Abstract:Aiming at the low measurement accuracy in paper defect detection due to the non-uniform image background and unobvious gray level feature, the paper suggested adopting Orthogonal Matching Pursuit(OMP) to conduct sparse decomposition by establishing a multi-scale redundant dictionary, the image background was compensated based on the background image and the different characteristics of paper defect image, thereby enhancing the paper defect characteristics. Experiment showed that this method could efficiently restructure and compensate the background image, highlight the paper defects with low gray level characteristic, eventually to improve the accuracy of paper defect detection．
查看全文  HTML  查看/发表评论  下载PDF阅读器