基于压缩感知的等效源近场声全息

耿胜远; 张学勇

文章摘要

耿胜远,张学勇.基于压缩感知的等效源近场声全息[J].安徽建筑大学学报,2021,29(4):77-82

基于压缩感知的等效源近场声全息

Near-field Acoustic Holography Technology Based on Compressed Sensing Theory and Equivalent Source Method

DOI：

中文关键词: 近场声全息压缩感知等效源法声场重建声源识别

英文关键词: near-field acoustic holography compression sensing equivalent source method sound field reconstruction source identification

基金项目:国家自然科学基金项目 (61471003)；国家重点研发计划项目（2018YFB2003801）；安徽省教育厅自然科学基金重点研究项目（KJ2020A0484）。

作者	单位
耿胜远	安徽建筑大学数理学院，安徽合肥 230601；安徽省建筑声环境重点实验室，安徽合肥 230601
张学勇	安徽建筑大学数理学院，安徽合肥 230601；安徽省建筑声环境重点实验室，安徽合肥 230601

摘要点击次数: 6081

全文下载次数: 5046

中文摘要:

压缩感知利用感知矩阵可从稀疏信号中高概率重建出原始信号。本文介绍一种压缩感知稀疏采样等效源近场声全息方法，对频率为 1500 Hz 的两个单极子声源和频率在 500 Hz、1000 Hz、2000Hz 时的六个单极子声源，通过采用 21×21 网格点（441 个）及 121 个随机数据采样点，分别进行声场重建与声源识别仿真，并与传统等效源近场声全息方法进行比较。结果表明，121 和 441 个采样点下的传统等效源法声场重建误差分别为 26.79% 和 45.21%，而基于本文提出的压缩感知等效源法的声场重建误差仅为 2.24% 和 0.92%，因而，具有更高的声场重建能力；在 121 个采样点下，压缩感知等效法声源识别图像分辨率优于传统等效源法，声源定位及识别精度高。

英文摘要:

Compressed Sensing is a technique that uses a perception matrix to reconstruct the original signal with high probability from sparse signals.Near-field acoustic holography method based on compressed sensing sparse sampling and equivalent source is presented.For two monopole sub-sound sources with frequency of 1500 Hz and six monopole sub-sound sources with frequency of 500 Hz 、1000 Hz and 2000 Hz ，by using 21×21 grid points （441） and 121 random data sampling points，sound field reconstruction and source identification simulation are carried out respectively and the comparison is made with the traditional equivalent source near field acoustic holography method.The results show that the reconstruction errors of the traditional equivalent source method under 121 and 441sampling points are 26.79% and 45.21%，respectively，while the errors of the proposed equivalent source method based on compressed sensing are only 2.24% and 0.92%，so it has a higher ability of sound field reconstruction.Under 121 sampling points，the resolution of sound source recognition image based on compressed sensing is better than that of traditional equivalent source method，and theaccuracy of sound source location and identification is high.

查看全文查看/发表评论下载PDF阅读器

关闭