摘要
红外无创血糖检测是一种不需收集血样进行人体血糖浓度测量的新方法。它不仅可以减少患者的痛苦 ,还能进行连续测量。由于无创血糖检测技术市场利益巨大是各国研究的热点 ,且技术相互之间保密 ,可参考的数据不多。常用的技术是利用一束一定波长的红外光传过人体部分血管区域 ,从光谱中提取相应的血糖浓度信息 ,但在测量精度和稳定性方面存在着问题。为了克服红外无创血糖检测中存在的难题 ,使微弱的光谱信号变化能正确的体现人体血糖浓度 ,作者根据朗伯—比尔定律原理 ,分析了人体葡萄糖的吸收谱 ,设计了多波长红外血糖检测传感器阵列 ,采用MOE和MADALINE整合神经网络方法建立了传感器阵列信号处理模型 ,使红外无创血糖检测的精度和稳定性得到了改善 。
The noninvasive blood glucose measurement is a novel technique. It shows a lots of benefits in markets, so the research of instruments on noninvasive blood glucose measurement is a focus. The technological details still keep in secret and lack of reference data . Although we can use this technique to measure the concentration of glucose in the blood noninvasively by single beam of infrared without obtaining a drop of blood and pain, but it remains some problems of accurate degree and stability to be resolved. According to the Lambert Beer Law, author designed the infrared multiwavelength sensor array and used integrated neural network of MOE and MADALINE to process array signals. It can overcome existing problems and get more exact blood glucose information from faint changes of spectral signals. The problems need to be solved are also discussed.
出处
《红外技术》
CSCD
北大核心
2003年第2期60-64,共5页
Infrared Technology
基金
国家自然科学基因资助 ( 60 0 710 2 9)
西安交通大学项目培植基金资助
甘肃省自然科学基金资助
