基于非接触式拉曼光谱分析人血与犬血的PCA-LDA鉴别方法

发布时间：2018-03-11 07:00

本文选题：人血　切入点：犬血　出处：《高等学校化学学报》2017年04期 　论文类型：期刊论文

【摘要】：将拉曼光谱分析法与数理统计方法有机结合,构建人血与犬血种属判别模型,实现了不同种属血液样本的高效无损鉴别.采用拉曼光谱的无损测试模式对血液样本进行测试,考察了抗凝管管材、聚焦位置及曝光时间等对血液样本拉曼光谱的影响,在激发波长为632.8 nm,光谱扫描范围为200~1800 cm-1,功率衰减率50%,曝光时间5 s及累加次数为2次的优化条件下,获得了无损检测条件下的血液样本拉曼光谱图.针对血液样本组分复杂、拉曼光谱信号基底背景高等问题,提出了基于小波变换去噪,进行分段多项式基线校正的预处理方法,有效解决了血液样本拉曼光谱谱图的高噪音和基线漂移问题.实验选择30例正常人血和33例比格犬血为样本训练集,5例正常人血和5例比格犬血为测试集,基于主成分分析法(PCA)联合线性判别法(LDA)模型,训练集分类正确率达到95.23%,盲测集分类正确率达90.00%.这种基于非接触式血液样本拉曼光谱和PCA-LDA判断模型的测试方法在进出口检验检疫等涉及血液无损鉴别的领域具有广泛的应用价值和前景.
[Abstract]:Combining the Raman spectrum analysis method with the mathematical statistics method, a discriminant model of human and dog blood species was constructed, and the high efficient nondestructive identification of the blood samples of different species was realized. The blood samples were tested by the non-destructive testing mode of Raman spectroscopy. The effects of anticoagulant tube material, focusing position and exposure time on Raman spectra of blood samples were investigated. Under the optimized conditions, the excitation wavelength was 632.8 nm, the spectrum scanning range was 200 ~ 1800 cm ~ (-1), the power attenuation rate was 50%, the exposure time was 5 s and the cumulative times were 2 times. The Raman spectra of blood samples under the condition of nondestructive testing are obtained. Aiming at the problems of complex components of blood samples and high background of Raman spectrum signals, a pretreatment method based on wavelet transform is proposed to carry out segmented polynomial baseline correction. The problems of high noise and baseline drift of Raman spectrum of blood samples were effectively solved. 30 normal human and 33 Beagle's blood were selected as sample training set. 5 normal human and 5 Beagle's blood were used as test set. Based on the principal component analysis (PCA) combined with linear discriminant method (LDA), The accuracy rate of training set classification is 95.23 and that of blind test set is 90.00. this method based on non-contact blood sample Raman spectrum and PCA-LDA judgment model has been used in import and export inspection and quarantine and other fields involving blood nondestructive identification. Extensive application value and prospect.
【作者单位】：重庆大学新型微纳器件与系统技术重点学科实验室;重庆大学化学化工学院;重庆大学光电工程学院微系统研究中心;重庆大学软件学院;
【基金】：国家自然科学基金(批准号:21375156) 重庆市科委基础科学与前沿技术重点项目(批准号:cstc2015jcyj BX0010) 国家“八六三”计划项目(批准号:2015AA021104)资助~~
【分类号】：R446.1;O657.37

【相似文献】