在日常活动中实时连续获取人体脉搏信号对心血管疾病诊断和医疗系统具有重要的临床意义。近年来，基于对聚合物多模光纤散斑传感技术日渐成熟，由于其固有的安全性、紧凑性、电隔离性以及低成本实现高灵敏度的传感特性，该技术在人体生理监测方面表现出极大的发展潜力。在本文中，我们开发了一套基于散斑图像分析的可穿戴设备，将塑料光纤（POF）集成在运动护腕上，用于人体脉搏信号的监测。本文对几种不同的光纤散斑图样处理算法进行了评估与比较，平方差之和（SSD）算法表现最佳，以不到100s的低计算时间获得了高精度结果，测量误差控制在4.9%左右。为了提高POF传感器的性能，在使用SSD算法的基础下选择了四种不同内径的POF进行进一步的实验探究与优化。测试结果表明，750 um内径的POF传感器具有最高的信噪比（34.96 dB）和低延时性（与ECG参考信号监测相比存在369.9 ms的延迟），同时将测量误差降低至3.7%。优化后获得的人体脉搏波信号能够被用来进一步的医学诊断，这对临床上心血管疾病的治疗以及家庭监测显示出潜在的应用价值。

The real-time acquisition of human pulse signals in daily life has clinical significance for the monitoring and diagnosis of cardiovascular diseases. In recent years, multimode plastic optical fiber speckle sensing technology has shown great potential for human physiological monitoring due to its inherent safety, compactness, electrical isolation, and low cost to achieve high sensitivity sensing characteristics. In this paper, the wearable device is designed based on speckle pattern analysis with plastic optical fiber (POF) integrated into a sports wrister for human pulse signal monitoring. Several different fiber speckle pattern processing algorithms were evaluated and compared, and the sum of squared differences method (SSD) performed the best, obtaining high accuracy results (the measurement error is controled at about 4.9%) with a low computational cost of less than 100s. To improve the performance of the sensor, POFs with different core diameters were selected for experimental investigation and optimization based on the SSD method. The results indicate that the POF sensor with 750 um core diameter has the highest signal-to-noise ratio (34.96 dB) and low-latency (Delay of 369.9 ms compared to reference measurement), while the measurement error was reduced to 3.7%. The optimized human pulse signal can be used for further medical diagnosis, which shows potential applications for the clinical treatment of cardiovascular diseases and home monitoring.