数学形态学滤波器的工程设计与实验研究

数学形态学是非线性数字信号处理学科的热门课题，将其分支——平结构元滤波用于工程设计时，结构元长度的选取是关键．为此，从检波理论出发，讨论形态学滤波器消除基线漂移的机理，分析其时域和频域特点，得到信号的衰减幅度由信号数字频率与结构元长度的比值唯一决定的性质，并归纳出一套平结构元长度选取标准。将其用于心电预处理，借助Matlab软件进行算法模拟并移植到DSP硬件平台，实验表明，该方法能有效并实时滤除基漂，符合后期处理需要。

Engineering Design and Experiment of Mathematical Morphology Filter

For flat structuring element filtering, a branch of mathematical morphology, length selection of structuring element is the key to its application to engineering design. Based on detection theory and time-frequency-domain analysis of the fundamentals of mathematical morphology filter, it is found that the ratio of the signal numerical frequency to length of structuring element determines the attenuation magnitude of signal. Then the criteria for the selection of flat structuring element is obtained and applied to preprocessing electrocardiograph (ECG). The arithmetic is simulated with Matlab software, and then transplanted to DSP hardware platform. It is shown that the method can effectively and instantly eliminate baseline drift of ECG, which fulfills the demand of final-period processing.