基于LabVIEW的心音身份识别系统

文中以LabVIEW 2012为开发平台,设计、实现了一种心音身份识别系统.该系统包括心音采集、数据处理和身份识别3个主要模块.心音采集模块利用自制传感器采集心音并去噪;数据处理模块完成心音特征提取和建立心音特征数据库;身份识别模块采用两种算法对心音特征数据进行分类识别,并使用决策层融合算法提高识别率.文中根据心音信号s1、s2的频谱特性和虚拟仪器的特点,提出了一种低频加强型梅尔倒谱系数和频域分段相关系数的特征提取算法,重点分析了基于矢量化欧式距离和最小相关距离分类识别方法.实际应用的结果证明该系统界面友好、操作方便、运算速度快、辨识效率高,具有一定的推广应用价值.

Heart Sound Identification System Based on LabVIEW

A heart sound identification system is designed and implemented based on LabVIEW. The system includes three main modules：signal collection, data processing and identification modules. The heart sound acquisition module collects heart sound and denoises by using self-made sensors. The data processing module completes the heart sound feature extraction and sets the up heart sound characteristic database. The identification module adopts two kinds of algorithm for classification of heart sound recognition feature data, and uses the decision level fusion algorithm to improve the recognition rate. Based on the spectral characteristics of heart sound signals of sl and s2, and the virtual instrument, this paper presents a feature extraction algorithm with a low frequency enhanced Mel cepstral coefficients and the frequency division coefficient. It can analyze the vector Euclidean distance and the minimum distance classification method. The practical application shows that the system has friendly interface, convenient operation, fast operation speed, and high identification efficiency.