基于太赫兹时域光谱的农药定性和定量分析若干关键技术研究

Terahertz time-domain spectroscopy (THz-TDS) is gaining more interests from various research fields due to its uniqueness. with THz-TDS,the 0.5-1.5 THz absorbance spectra of pesticide solutions with the concentration in the level of several ug/mls are analyzed, as well as their relationship to the solution concentrations.And a new method based on the THz-TDS technique and the combination of partial least square(PLS) and least-squares support vector machine(LS-SVM) modeling mehods is proposed for the detection of pesticide solutions. According to the model building and prediction preocesses, sources influencing the measurement precision are analyzed theoretically as well. The signal to noise ratio of the THz time domain spectroscopy was in-depth analyzed by Wavelet. Multi-layer decomposition and reconstruction of discrete wavelet,soft-threshold filtering and hard-threshold filtering of wavelet transform were applied respectively in THz time domain spectroscopy, and then the signal to noise ratio of the THz time domain spectroscopy was detail analyzed and improved. Least Squares Support Vector Machines, Naive Bayes and Back Propagation Artificial Neural Network were applied to achieve Multi-class classification of these different kinds of pesticides, and the classification results of three algorithms were analyzed in detail. Based on the database of THz spectroscopy and principal component analysis algorihm, recognition functions were realized. Eventually, a set of label-free detection prototype device based on terahertz time-domain spectroscopy will be developed.

太赫兹时域光谱（THz-TDS）技术由于其独特的光谱特性，正受到各领域学者越来越多的关注。本项目将利用THz-TDS技术分析浓度为ug/ml级的农药溶液在0.5-1.5THz的吸收谱特性及其与溶液浓度的关系，提出采用THz-TDS与全光谱偏最小二乘-最小二乘支持向量机相结合的方法，从理论上分析影响测量精度的误差来源；采用小波变换深入研究了THz时域光谱的信噪比。将离散小波多层分解和重构技术、小波变换的软阀值滤波和硬阀值滤波技术应用于THz时域光谱的噪声去除，提高THz时域光谱的信噪比；分别采用LS-SVM, Naive-Bayes 和 BP-ANN建模，实现不同品种农药的分类识别。同时，采用离子群算法和遗传优化算法优化了LS-SVM参数的求解；在建立物质THz吸收谱数据库的基础上，采用主成分分析法实现物质的在线识别检测；最终开发一套基于THz-TDS的农药残留无标记检测原型装置。

太赫兹时域光谱（THz-TDS）技术由于其独特的光谱特性，正受到各领域学者越来越多的关注。本项目利用THz-TDS技术分析浓度为ug/ml级的农药溶液在0.5-1.5THz的吸收谱特性及其与溶液浓度的关系，提出采用THz-TDS与全光谱偏最小二乘-最小二乘支持向量机相结合的方法，从理论上分析影响测量精度的误差来源；采用小波变换深入研究了THz时域光谱的信噪比。将离散小波多层分解和重构技术、小波变换的软阀值滤波和硬阀值滤波技术应用于THz时域光谱的噪声去除，提高THz时域光谱的信噪比；. 采用固体密度泛函仿真的方法对物质的吸收峰来源进行了研究，并与分子密度泛函仿真方法相比较，结合理论计算与实验结果，对其振动来源进行了分析指认。在计算获取过氧化苯甲酰和聚乙烯混合物、过氧化苯甲酰和面粉混合物THz 吸收谱和折射率谱的基础上，使用小波滤波、一阶导数和二阶导数进行预处理，而后分别采用PLS和KPLS 对原始吸收谱、一阶导数谱和二阶导数谱回归分析建模，并给出了详细性能评价。. 深入研究了基于THz 光谱的物质分类和识别检测。在检测获取四种常见绿茶太赫兹光谱的基础上，分别采用LS-SVM，Naive-Bayes 和BP-ANN 建模，实现不同品种绿茶的分类识别。同时，采用粒子群算法和遗传算法优化了LS-SVM 参数的求解，并对其实验结果进行了分析比较。. 分别采用LS-SVM, Naive-Bayes 和 BP-ANN建模，实现不同品种农药的分类识别。同时，采用离子群算法和遗传优化算法优化了LS-SVM参数的求解；在建立物质THz吸收谱数据库的基础上，采用主成分分析法实现物质的在线识别检测；最终开发一套基于THz-TDS的农药残留无标记检测原型装置。