改进的经验模态分解法去除脉搏波基线漂移

脉搏波信号中含有丰富的人体生理病理信息，然而基线漂移的存在严重影响人体生理参数的准确提取，需要予以去除。针对传统经验模态分解（EMD）方法在去除基线漂移时，通过经验来确定基线漂移信号所在的固有模态函数阶数，导致去噪性能下降。根据EMD分解过程中固有模态函数的特性，提出基于过零率检测的方法进行算法改进，通过计算每阶固有模态函数的总过零率，设定阈值来确定基线漂移阶数。仿真实验结果表明：该改进算法科学严谨，能够有效地去除脉搏波基线漂移，提高了信噪比，有利于提高基于光电容积脉搏波提取血压、血氧等人体生理参数的精度。

Improved Empirical Mode Decomposition to Remove the Baseline Drift of Pulse Waves

Pulse wave signal contains abundant physiological and pathological information, however, the existence of baseline drift seriously affects the accurate extraction of human physiological parameters and needs to be removed. For the traditional empirical modal decomposition (EMD) method, when the baseline drift is removed, The intrinsic mode function order of the baseline drift signal is determined by the experience, which leads to the decrease of the denoising performance. Based on the characteristics of EMD decomposition process, this paper proposes a method based on zero-crossing rate detection to improve the baseline drift order by calculating the zero-crossing rate of each layer''s intrinsic modal function and setting a reasonable threshold. The experimental results show that the algorithm proposed in this paper is scientific and reasonable, can effectively remove the baseline drift of pulse wave, improve signal-to-noise ratio and improves the accuracy of photoelectric pulse physiological parameters of blood pressure, blood oxygen wave extraction.