基于六阵列传感器的全自动脉搏信号采集系统  

Automatic pulse signal acquisition system based on six-array sensors

作　　者：田恒屹 黄璐琦[2] 咸楠星 张宇飞 李迎 赵宇平[2] TIAN Hengyi;HUANG Luqi;XIAN Nanxing;ZHANG Yufei;LI Ying;ZHAO Yuping(Experimental Research Center,China Academy of Chinese Medical Sciences,Beijing 100700,China;China Academy of Chinese Medical Sciences,Beijing 100700,China;Tianjin Construction Machinery Research Institute Co Ltd,Tianjin 300409,China)

机构地区：[1]中国中医科学院医学实验中心,北京100700 [2]中国中医科学院,北京100700 [3]天津工程机械研究院有限公司,天津300409

出　　处：《传感器与微系统》2025年第1期83-87,共5页Transducer and Microsystem Technologies

基　　金：国家中医药创新团队项目(ZYYCXTD-D-202205)。

摘　　要：脉诊是中医临床诊断的关键步骤之一。在中医四诊信息客观化研究中,针对传统脉诊仪采集时间长及需要医生辅助定位的问题,基于中医“三部九候”理论,开发了一种全流程自动化的脉搏信号采集系统。系统利用计算机视觉与深度学习模型准确地定位“寸关尺”位置,并集成了机械加压控制系统,实现在不同压力下对双手脉搏信号的同步采集。性能测试结果显示:该系统能够准确检测脉搏位置,mAP@0.5值高达99.9%。此外,系统的采集准确率高,符合无人化采集的需求。该系统有望用于医院的预诊环节,能快速、便捷地收集患者的脉诊客观化信息,显著缩短患者的检测时间,并减轻医生使用脉诊仪的工作压力。Pulse-taking diagnosis is a crucial step in traditional Chinese medicine(TCM)diagnostics.In research on objectification of information of four diagnoses in TCM,aiming at the problem of long collection times traditional devices and needs physician to assist positioning,an automated pulse signal acquisition system is developed based on“three regions and nine divisions”theory of TCM.The system accurately locates pulse points using computer vision and deep learning model,and integrates a mechanical loading control system to realize simultaneous,dualwrist pulse signal acquisition under various pressures.Performance tests result show that the system can detect pulse position accurately,with a [e-mail] system has high accuracy of acquisition,which is suitable for unmanned acquisition and pressure in hospitals.It can collect pulse diagnosis objectifies information quickly and easily,significantly shorten testing time and reduce physician work.

关 键 词：自动化 机器视觉 中医脉诊 脉象采集 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]