Computer Engineering and Applications ›› 2011, Vol. 47 ›› Issue (25): 186-189.

• 图形、图像、模式识别 • Previous Articles     Next Articles

Improved model of image block compressed sensing

LI Yunhua   

  1. School of Electronic Information Engineering,Nantong University,Nantong,Jiangsu 226007南通大学 电子信息学院,江苏 南通 226007
    School of Electronic Information Engineering,Nantong University,Nantong,Jiangsu 226007,China
    ,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-09-01 Published:2011-09-01

一种改进的图像分块压缩感知模型

李蕴华   

  1. 南通大学 电子信息学院,江苏 南通 226007

Abstract: The technique of block compressed sensing which is used in image acquisition can solve the problem of large amount of computing which exits in image construction when the traditional compressed sensing is used.But using block compressed sensing,the constructed image quality is degraded.An algorithm of improved block compressed sensing is proposed.By designing a weighting scheme for the sampling matrix,the reconstruction accuracy of low-frequency part of the image can be increased,and so the quality of image is improved.And by altering adaptively the measurements number according to the different texture features of the image block,the number of measurements will be much fewer under the premise of image quality.The experiment proves the efficiency of the algorithm.

Key words: compressed sensing, image construction, sparse representation, signal sampling

摘要: 分块压缩感知用于图像获取可以解决传统压缩感知在重构时运算量大的问题,但是运用分块压缩感知却使重构图像的质量有所降低。提出了一种改进的图像分块压缩感知算法。该算法通过对观测矩阵加权,保证了图像低频部分在重构时获得更大的精度,提高了图像的质量。另外,算法根据各图像块不同纹理复杂性,自适应地改变观测值数目,使得在保证图像质量的前提下,重构所需的总观测值数目更少。实验证明了该算法的有效性。

关键词: 压缩感知, 图像重构, 稀疏表示, 信号采样