摘要
中红外光谱法不使用任何试剂检测血糖浓度,但由于精度不够高,尚未应用于临床检测。本文从中红外光谱法测量血糖浓度的原理、技术出发,结合实验研究结果,从选取最佳研究波段、光谱处理方法、消除血液中其他成分干扰及改进系统硬件等四方面分析了当前中红外光谱法测血糖可以提高测量和预测精度的途径。针对四条途径,本文分别进行了以下实验:(1)严格确定在中红外区对葡萄糖浓度变化最灵敏的区域;(2)应用遗传算法筛选波长;(3)光谱正规化以提高仪器对样品测量的再现性;(4)使用高能量的CO2激光器作为光源,实验测量结果提高至接近或达到临床检测的可接受水平。
Mid-infrared(MIR) spectroscopy is applicable to blood glucose sensing without using any reagent, however, due to a result of inadequate accuracy, till now this method has not been used in clinical detection. The principle and key technologies of blood glucose sensing by MIR spectroscopy are presented in this paper. Along with our experimental results, the paper analyzes ways to enhance measurement accuracy and prediction accuracy by the following four methods: selection of optimized spectral region; application of spectra data processing method; elimination of the interference with other components in the blood, and promotion in system hardware. According to these four improving methods, we designed four experiments, i. e. , strict determination of the region where glucose concentration changes most sensitively in MIR, application of genetic algorithm for wavelength selection, normalization of spectra for the purpose of enhancing measuring reproduction, and utilization of CO2 laser as light source. The results show that the measurement accuracy of blood glucose concentration is enhanced almost to a clinical detection level.
出处
《生物医学工程学杂志》
EI
CAS
CSCD
北大核心
2006年第3期688-691,共4页
Journal of Biomedical Engineering
基金
国家自然科学基金资助项目(30170261)
天津市自然科学基金资助项目(023800411)
关键词
中红外
血糖
衰减全反射
偏最小二乘法
遗传算法
Min-infrared(MIR) Blood glucose Attenuated total reflection (ATR) Partial least squares (PLS) Genetic algorithms