摘要
压缩感知理论基于信号的稀疏性,压缩感知技术在采集信号的同时,实现数据的压缩处理,能够显著减少传输过程中纸病图像的数据量。结合纸病图像的特点,在研究了纸病图像稀疏性的基础上确定了测量矩阵,完成了计算机PC重建时的重构算法。通过仿真实验,验证了不同的稀疏基和采样率对纸病图像重构质量的影响。结果表明,利用压缩感知技术,纸病图像数据的传输量只有原来的30%~40%,并且重构的图像质量也较好,能够在一定程度上提高造纸生产线上纸病检测的速度。
With the increase of machine＇s speed and paper width in the paper industry, the amount of image data acquisition and transmission is larger, the poor real-time of on-line detection has been a bottleneck in the paper disease detection. Based on sparsity of the signal, compressed sensing can realize the compression of the data while the data are collecting, and significantly reduce the data amount in the transmission. Com- bining with the characteristics of paper disease images, this paper determined the measurement matrix based on the study of the sparsity, and obtained reconstruction algorithm on PC. Through the simulation experiment, the impact of different sparse matrix and different sampling rate on the quality of the disease image reconstruction was verified. The result showed that, using this technology, data transmission amount reduced 60% - 70% and the quality of the reconstructed image was good, this could improve the speed of paper disease detection to a certain extent.
出处
《中国造纸学报》
CAS
CSCD
北大核心
2015年第3期51-56,共6页
Transactions of China Pulp and Paper
基金
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
关键词
压缩感知
纸病图像
数据实时采集和传输
compressed sensing
paper disease image
data real-time acquisition and transmission
作者简介
周强，男，1969年生；博士，教授；主要研究方向：智能信息处理技术。
通信联系人：王志强，E-mail：1098187867@qq．com
