用EMD和小波消噪的加速度信号压缩重构新方法.pdfVIP

2015 年第24 卷第 11 期 htφ:// 计算机系统应用 用 EMD 和小波消噪的加速度信号压缩重构新方法① 吴建宁，徐海东 (福建师范大学数学与计算机科学学院，福州 350007) 摘 要:针对噪声破坏加速度信号稀疏性、降低其压缩感知重构算法性能问题，提出了一种用经验模态分解 (EMD)和小波分析联合消噪的加速度信号压缩重构新方法.该方法首先采用 EMD 和小波阑值联合消噪方法对加 速度信号消噪处理，保持加速度信号内在稀疏性;然后基于压缩感知理论和加速度信号块结构信息，采用块稀疏 贝叶斯学习算法以高概率重构原始加速度信号.采用USC-HAD人体日常行为数据库中的加速度信号验证本文方 法的有效性.实验结果表明，本文所提方法的信噪比和均方根误差明显优于未经消噪处理的压缩感知重构算法， 能够有效抑制加速度信号噪声，增大加速度信号稀疏度，提高加速度信号压缩重构算法性能. 关键词:去噪;加速度信号;经验模态分解(EMD); 小波阐值;压缩感知(CS) New Method for Compression and Reconstruction of Acceleration Signal Using EMD and Wavelet Denoising WU Jian-Ning, XU Hai-Dong (College ofMathmatics and Computer Science, Fujian Normal Universi纱 Fuzhou 350007, China) Abstract: In order to solve 优 issue of the deterioration of reconstruction performance in compressed sensing (CS) for acceleration signal with poor sparsity that is produced by noise，也is paper proposed a new method for perfect reconstruction of acceleration signal by using empirical mode decomposition (EMD) and wavelet denoising. Its basic idea is 仕蹦出e best sp缸sity of acceleration signal is firstly gained by using EMD and wavelet threshold denoising me也od. And 吐1钮， considering CS and acceleration signal with block structure, block spru回 Bayesian learning algorithm is applied to perfect1y reconstruction original acceleration signa1. The acceleration signal 企om human activity dataset USC-HAD is selected to test the effectiveness of the proposed method. The experimental results show 也at when compared to traditional CS algori也m without pre-denoising processing，仕le proposed me也od can obtain the best value of signal-to-noise ratio and root mean square eηor. Also , these results sugg臼ted that our proposed method has the superior ability of denoise for gaining best sparsity of acceleration signal, which si伊ificant1y improve the reconstruction pe由rmance of CS. Key words: denoising; acceleration signal; empirical mode decomposition