中文摘要：

无创血液成分测量作为一种急需的医疗检验技术具有十分重要的临床应用价值,由于未能消除个体差异和测量环境等因素的影响,该技术难以进入临床应用领域.动态光谱法从理论上解决了上述问题.基于动态光谱理论首次尝试对人体血液中多种蛋白含量的无创测量,采集110名志愿者指端透射光谱信息,并以血液成分含量生化分析值为参考,建立血液中多种蛋白含量与动态光谱数据的偏最小二乘法（PLS）校正模型.对血红蛋白、总蛋白和白蛋白含量校正模型的相关系数分别为0.999 4、0.922 1和0.956 0,预测集的平均相对误差为3.77%、3.49%和3.40%.结果表明,应用动态光谱对血液蛋白含量进行无创测量的方法可获得较高的精度,也为近红外光谱法对血液成分的无创测量提供了一条新的路径.

英文摘要：

Non-invasive determination of blood components concentration,as one of the medical analysis technique,has attracted much interest in recent years,but it has not yet been applied to the clinical field due to the effect of individual discrepancy and complicated measurement condition.To solve these problems,a non-invasive（NI）in vivo measurement technique of human blood proteins concentration using near-infrared dynamic spec-trum（DS）method was put forward.In vivo measurements were carried out on 110 volunteers,and biochemical ex-aminations of blood components contents,including hemoglobin（HGB）,total protein（TP）and albumin（ALB）were conducted simultaneously to establish partial least squares（PLS） calibration models.Correlation coefficients between calibration sets and prediction sets were 0.999 4,0.922 1 and 0.956 0 respectively,and the average relative errors of prediction sets were 3.77%,3.49% and 3.40% respectively.The results show that the dynamic spectrum analysis can achieve high accuracy and has provided a promising NIR approach to the non-invasive measurement of human blood components.