基于Mel频率倒谱系数、动态时间归准的计算机肺音分析系统在健康儿童中的应用

目的 探索基于Mel频率倒谱系数（MFCC）、动态时间归准（DTW）设计计算机肺音分析系统，对儿童肺音进行识别匹配，验证该肺音分析系统的有效性。 方法 设计计算机肺音分析系统，采用Lung Sounds Reference Guide的标准肺音（包括正常、异常肺音）及采集典型的湿啰音、鼾音、哮鸣音各1例（分别来自厦门大学附属第一医院儿科3例患儿）作为参考肺音；通过自行研制的便携式电子听诊器，对29例健康儿童肺音进行采集作为测试肺音。通过人工识别有效肺音片段及MFCC对测试肺音进行特征提取，而后对经MFCC处理后的肺音数据进行DTW模板匹配，最终得出识别结果。结果 实验获得参考肺音39个，测试肺音58个。实验共进行了58次肺音识别，正确识别52次，错误识别6次，识别正确率为89.7%。结论 基于MFCC、DTW的计算机肺音分析系统对健康儿童肺音（正常肺音）的识别是有效的。

Research on measuring time constant of NANMAC thermocouple

The theory for measuring the time constant of thermocouple was introduced, and the method for measuring the time constant of NANMAC thermocouple by using dynamic calibration system of transient surface temperature sensor was proposed. In this system, static and dynamic calibrations were conducted for infrared detectors and thermocouples, and then both temperature-time curves were obtained. Since the frequency response of infrared detector is superior to that of calibrat- ed thermocouple, the values measured by infrared detectors are taken as true values. Through dividing the values measured with thermocouples by those with infrared detectors, a normalized curve was obtained, based on which the time constant of thermocouple was measured. With this method, the experiments were carried out with NANMAC thermocouple to obtain its time constant. The results show that the method for measuring the time constant is feasible and the dynamic calibration of thermocouples can be achieved at microsecond and millisecond level. This research has a certain reference value for research and application of NANMAC thermocouple temperature sensor.