中文摘要：

人体脉搏信号中隐藏了重要的生理和病理信息，但由于采集到的脉搏波中混入了不同的噪声，严重影响了波形特征点的准确识别。因此在比较了傅里叶变换与小波变换滤波效果的基础上，选用小波变换的方法对脉搏波进行基线漂移和高频噪声的去除，极大地提高了脉搏波信号的信噪比，接着利用脉搏波的时空特点提出了一种识别6个关键特征点的快速算法。最后结合Matlab与Java语言各自的特点，采用混合编程的方式设计了实现上述算法的安卓客户端软件。经实际测试．脉搏波特征点的识别速度和识别准确率均达到了预期效果，具有很好的实用价值。

英文摘要：

Human pulse signals include important physiological and pathological information; however, a signal mixed different noises is usually contained in the waveforms during acquisition, and it will impact on the accurate recognition of feature points. Based on the comparison of the filtering effects of Fourier transform and wavelet transform, the latter is selected to remove the baseline drift and filter high frequency noise in the pulse waves. Then a fast algorithm of recognition of six key feature points is proposed. Finally, Matlab and Java mixing programming is designed to realize the above algorithm on the Android client software. The practical tests illustrate that both in the speed of feature point recognition and the recognition accuracy the algorithm achieves the desired effect, hence it will have a good practical value.