[1]孙润润.基于稀疏度区间的变步长最优子空间追踪算法[J].计算机技术与发展,2019,29(02):48-53.[doi:10．3969/j．issn．1673－629X．2019．02．010]
　SUN Runrun.Variable Step Size Optimal Subspace Pursuit Algorithm Based on Sparsity Interval[J].,2019,29(02):48-53.[doi:10．3969/j．issn．1673－629X．2019．02．010]
点击复制

基于稀疏度区间的变步长最优子空间追踪算法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
29
期数:
2019年02期
页码:
48-53
栏目:
智能、算法、系统工程
出版日期:
2019-02-10

文章信息/Info

Title:
Variable Step Size Optimal Subspace Pursuit Algorithm Based on Sparsity Interval
文章编号:
1673－629X(2019)02－0048－06
作者:
孙润润
安徽理工大学 计算机科学与工程学院，安徽 淮南 232001
Author(s):
SUN Run－run
School of Computer Science and Engineering，Anhui University of Science and Technology，Huainan 232001，China
关键词:
贪婪算法; 稀疏自适应; 变步长; 最优子空间
Keywords:
greedy algorithm; sparse adaptivity; variable step size; optimal subspace
分类号:
TP391
DOI:
10．3969/j．issn．1673－629X．2019．02．010
摘要:
针对压缩感知逐步最优子空间算法(stepwise optimal subspace pursuit，SOSP)在迭代过程中使用固定步长和稀疏度未知的问题，提出一种基于稀疏度区间的变步长最优子空间追踪算法。该算法首先是在 SOSP 最优扩增缩减方法的基础之上根据匹配测试公式获取初始固定步长，再使用增删迭代公式，在假定支撑集的增加过程中根据信号残差能量的变化判断稀疏度的区间。最后在稀疏度区间中使用黄金分割法(golden ratio)逐渐减小区间得出信号稀疏度，并在区间分割的过程中逐渐删减假定支撑集中多余的元素，最终重构原始稀疏信号。实验结果表明，对于一维信号重建实验，在不同的稀疏度和测量值下，与同类算法相比，该算法都能够很好地重建原始信号且重构误差小。对于二维图像重建实验，图像重建精度较高且有很好的视觉效果。
Abstract:
In order to solve the problem of fixed step in the iterative process and unknown sparse signal in stepwise optimal subspace pursuit (SOSP)，we propose a variable step size optimal subspace pursuit algorithm based on sparsity interval． Firstly，based on SOSP，a method of optimal expansion and reduction，the initial fixed step size is obtained according to the matching test formula． Then，the iterative formulas about addition and deletion is used to judge the sparsity interval according to the change of signal residual energy in the increase of assumed support set． Finally，the golden ratio method is used to gradually reduce the interval for getting signal sparsity in the sparsity degree interval． In the process of interval segmentation，it deletes the unwanted elements in assumed support set and reconstructs the original sparse signal． The experiment shows that for one dimensional signal reconstruction experiment，the proposed algorithm is of little reconstruction error and reconstructs the original signal well under the different sparsity and measured values compared with the different algorithm． For two－dimensional image reconstruction experiment，it has high accuracy of image reconstruction with great visual effect．

相似文献/References:

[1]宋杰 程家兴 许中卫 周瑛.一种改进的贪婪式覆盖算法[J].计算机技术与发展,2006,(08):113.
　SONG Jie,CHENG Jia-xing,XU Zhong-wei,et al.An Improved Greedy Covering Algorithm[J].,2006,(02):113.
[2]石曼曼,李雷.基于分段可调节OMP算法的图像压缩感知算法[J].计算机技术与发展,2016,26(11):14.
　SHI Man-man,LI Lei. An Image Compressed Sensing Algorithm Based on Novel Stagewise Regulation OMP Algorithm[J].,2016,26(02):14.
[3]孙冬冬,杨龙祥. 基于软件定义的未来网络节能算法[J].计算机技术与发展,2017,27(03):70.
　SUN Dong-dong,YANG Long-xiang. Future Network Energy Saving Algorithm Based on Software Definition[J].,2017,27(02):70.
[4]刘芸菲,陆 奎.基于贝叶斯网络的电网事故预警模型[J].计算机技术与发展,2019,29(08):152.[doi:10. 3969 / j. issn. 1673-629X. 2019. 08. 029]
　LIU Yun-fei,LU Kui.Early Warning Model of Power Grid Accident Based on Bayesian Network[J].,2019,29(02):152.[doi:10. 3969 / j. issn. 1673-629X. 2019. 08. 029]

更新日期/Last Update: 2019-02-10