Hao Zhenhua, Cui Ziqiang, Yue Shihong, Wang Huaxiang
Tianjin Key Laboratory of Process Measurement and Control, School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, People's Republic of China.
Rev Sci Instrum. 2018 Jun;89(6):064702. doi: 10.1063/1.5021058.
As an important means in electrical impedance tomography (EIT), multi-frequency phase-sensitive demodulation (PSD) can be viewed as a matched filter for measurement signals and as an optimal linear filter in the case of Gaussian-type noise. However, the additive noise usually possesses impulsive noise characteristics, so it is a challenging task to reduce the impulsive noise in multi-frequency PSD effectively. In this paper, an approach for impulsive noise reduction in multi-frequency PSD of EIT is presented. Instead of linear filters, a singular value decomposition filter is employed as the pre-stage filtering module prior to PSD, which has advantages of zero phase shift, little distortion, and a high signal-to-noise ratio (SNR) in digital signal processing. Simulation and experimental results demonstrated that the proposed method can effectively eliminate the influence of impulsive noise in multi-frequency PSD, and it was capable of achieving a higher SNR and smaller demodulation error.
作为电阻抗断层成像（EIT）中的一种重要手段，多频相敏解调（PSD）可被视为测量信号的匹配滤波器，并且在高斯型噪声情况下是一种最优线性滤波器。然而，加性噪声通常具有脉冲噪声特性，因此有效降低多频PSD中的脉冲噪声是一项具有挑战性的任务。本文提出了一种用于EIT多频PSD中脉冲噪声降低的方法。在PSD之前，采用奇异值分解滤波器作为前级滤波模块，而不是线性滤波器，该滤波器在数字信号处理中具有零相位偏移、失真小和高信噪比（SNR）的优点。仿真和实验结果表明，所提出的方法能够有效消除多频PSD中脉冲噪声的影响，并且能够实现更高的SNR和更小的解调误差。
