基于LIBS的山药饮片产地溯源研究

山药为薯蓣科植物薯蓣的根茎,其中的多糖、多酚、皂苷、黏蛋白和维生素C等成分使山药具有抗肿瘤、抗氧化、抗炎症、降血糖和降血脂等作用。不同产地的山药由于生长条件存在差异,致使药用成分含量显著不同,结合独特的炮制工艺,进而导致市场价格差别大,所以山药饮片的产地识别至关重要。为对山药饮片进行产地溯源,本文基于激光诱导击穿光谱(LIBS)技术提出多元散射矫正-改进遗传算法-支持向量机(MSC-IGA-SVM)模型对山药产地进行精确识别。使用八个不同产地的山药饮片进行LIBS实验,八种产地的山药饮片磨粉过筛后制成粉末压片,通过采集山药饮片的LIBS光谱,分别使用单一分类器与使用光谱预处理、特征提取及模式识别算法的模型对光谱的识别结果进行对比。将光谱信号按2∶1的比例划分为训练集和测试集,使用5次交叉验证K-邻近算法(KNN)模型的测试集准确率作为预处理参数优化的评价指标。各类药材的平均光谱整体趋势一致,所含谱峰基本相同,但因产地不同导致峰值强度各不相同,道地山药对一些金属元素(K, Na, Ca, Mg, Al)的富集能力大于非道地产区山药,其中,K元素特征谱线(769.90 nm)的峰值最高,即...     

Randomized Preprocessing of Homogeneous Linear Systems of Equations

Our randomized preprocessing enables pivoting-free and orthogonalization-free solution of homogeneous linear systems of equations. In the case of Toeplitz inputs, we decrease the estimated solution time from quadratic to nearly linear, and our tests show dramatic decrease of the CPU time as well. We prove numerical stability of our approach and extend it to solving nonsingular linear systems, inversion and generalized (Moore-Penrose) inversion of general and structured matrices by means of Newton’s iteration, approximation of a matrix by a nearby matrix that has a smaller rank or a smaller displacement rank, matrix eigen-solving, and root-finding for polynomial and secular equations and for polynomial systems of equations. Some by-products and extensions of our study can be of independent technical intersest, e.g., our extensions of the Sherman-Morrison-Woodbury formula for matrix inversion, our estimates for the condition number of randomized matrix products, and preprocessing via augmentation.     

Chemometric assisted simultaneous spectrophotometric determination of four-component nasal solutions with a reduced number of calibration samples

The use of multivariate spectrophotometric calibration for the simultaneous determination of three active components and one excipient in nasal solutions is presented. The resolution of four-component mixtures of phenylephrine, diphenhydramine, naphazoline and methylparaben in a matrix of excipients has been accomplished by using partial least-squares (PLS-1) and a variant of the so-called hybrid linear analysis (HLA) named net analyte preprocessing (NAP). Notwithstanding the presence of a large number of components and their high degree of spectral overlap, they have been rapidly and simultaneously determined with high accuracy and precision, with no interference, and without resorting to extraction procedures using non-aqueous solvents. A simple and fast method for wavelength selection in the calibration step is used, based on the minimisation of the predicted error sum of squares (PRESS) calculated as a function of a moving spectral window. The use of calibration designs of reduced size has been attempted. Satisfactory results were obtained when the number of calibration samples was reduced from 25 (full central composite) to 17 (fractional central composite) using the net analyte-based NAP method.     

应用于微靶装配的显微视觉检测技术

针对如何实现微小尺度（μm～mm）靶零件装配过程中的在线检测，研制了应用于微靶装配的显微视觉检测系统，并对系统的构成与显微成像、图像预处理、特征检测等关键图像处理技术进行了详细论述。提出了针对微靶零件的新型多重滤波除噪算法和几何特征检测算法，并以铝台阶、柱腔等微靶零件为应用实例开展了实验研究。结果表明，该系统采用的算法能有效、快速、准确地对靶零件进行几何特征和实时位置检测，在检测视场为3mm时，其检测精度≤3μm，角度检测精度≤0．1°，适用于ICF微靶等微型器件的微装配。     

无网格MPM法三维前处理系统设计

无网格法在解决爆炸冲击、裂纹扩展、高速碰撞以及大变形等问题时,由于与网格无关,能够避免有限元法因网格畸变计算精度降低或计算终止的矛盾,是目前求解固体力学问题的有效方法之一,也是近年来研究的热点。前处理对于无网格法数值计算是非常重要的环节,基于面向对象程序设计思想,应用C 和OpenGL开发了三维图形交互方式的MPM法前处理系统,提出了应用递归四叉树法、扫描表示法和CSG法划分任意形状非等间距的MPM法质点单元,为建立通用的无网格前处理平台提出了有效的解决方案。     

A subspace fitting algorithm of acoustic vector sensor array and corresponding matrix pre-filter design

In order to ease the pass-band response distortion of the matrix pre-filter,a simple approach for designing matrix spatial filter is proposed,which minimizes the sum of the k maximal distortion norm（k is the number of the constraint points）within the pass-band,while constraining the filter response within the stop-band.Considering the costly amount of calculation of the high-resolution methods,an algorithm with small amount of calculation based on matrix pre-filtering and subspace fitting using acoustic vector array（MF-VSSF）is proposed.Through joint processing of signal subspace of both pressure and particle velocity,the pre-filtering matrix and the signal subspace is decreased to M-dimensional（M is the number of array-element）,hence reduces the time-consumption of the matrix pre-filter design and DOA searching.Simulation results show that,the method offers the same performance as MUSIC with pre-filtering,but has much lesser amount of calculation.Moreover,the designed prefilter can efficiently suppress the interference in the stop-band and improve the estimation and resolution performance of successive DOA estimators.     

The accuracy and precision of the advanced Poisson dead‐time correction and its importance for multivariate analysis of high mass resolution ToF‐SIMS data

Time‐of‐flight secondary ion mass spectrometry (ToF‐SIMS) data collected in single ion counting mode suffers from dead‐time effects that lead to potentially confusing artefacts when common multivariate analysis (MVA) methods are applied to the data. These artefacts can be eliminated by applying an advanced Poisson dead‐time correction that accounts for the signal intensity in the dead‐time window preceding each time channel. Because this correction is nonlinear, it changes the noise distribution in the data. In this work, the accuracy of this dead‐time correction and the noise characteristics of the corrected data have been analysed for spectra with small numbers of primary ion pulses. A simple but accurate equation for estimating the standard deviation in the advanced dead‐time corrected data has been developed. Based on these results, a scaling procedure to enable successful MVA of advanced dead‐time corrected ToF‐SIMS data has been developed. The improvements made possible by using the advanced dead‐time correction and our recommended scaling are presented for principal components analysis of a ToF‐SIMS image of aerosol particles on polytetrafluoroethylene. Recommendations are made for using the advanced dead time correction and scaling ToF‐SIMS data in order optimize the outcomes of MVA. Copyright © 2014 John Wiley & Sons, Ltd.     

基于近红外光谱的SG-MSC-MC-UVE-PLS算法在全血血红蛋白浓度检测中的应用

为提高全血血红蛋白浓度预测模型的预测精度,基于近红外光谱分析,首先对原始全血透射光谱数据分别进行均值中心化、标准化、标准正态变量变换(SNV)、多元散射校正(MSC)以及Savitzky-Golay（SG）卷积平滑结合MSC的预处理操作,最终选择预处理效果最好的SG-MSC方法作为数据预处理方法,其最大相关系数达到0.944 1。对SG平滑的平滑窗口宽度进行讨论,找出平滑效果最好的窗口宽度为27。数据预处理消除了全血吸收光谱的基线失真,提高了全血吸收光谱数据的信噪比。将190个样本（190个血红蛋白浓度对应的透射光谱数据）分为具有相近血红蛋白浓度分布的校正集和测试集,其中校正集为143个样本（对应血红蛋白浓度分布为10.6～17.3 g·dL-1）,测试集为47个样本（对应血红蛋白浓度分布为10.3～17.3 g·dL-1）,确保建立模型的适用性。对校正集数据预处理后利用蒙特卡洛无信息变量消除（MC-UVE）方法对其进行波长变量选择,剔除含信息量少的波长点,提高含信息量多的波长占比。设置蒙特卡洛迭代次数为1 000,最终从全血吸收光谱的700个波长变量中筛选出191个波长变量用于建立全血血红蛋白浓度偏最小二乘（PLS）回归模型。对比分析原始全血透射光谱全谱PLS模型、原始全血吸收光谱全谱PLS模型、预处理全血吸收光谱全谱PLS模型、SG-MSC-MC-UVE-PLS模型以及已有二阶导数PLS模型的模型效果,表明基于SG-MSC-MC-UVE-PLS算法的全血血红蛋白浓度预测模型效果较其他模型效果更优,预测相关系数由0.676 3提高到0.979 1,预测集均方根误差由0.898 1减小到0.220 3,最大绝对误差由2.426 1减小到0.411 2。同时,利用MC-UVE方法进行波长变量选择,在保证预测精度的前提下,筛选出建模的波长个数更少,有利于提高模型计算效率。研究结果表明,SG-MSC-MC-UVE-PLS方法能够提高全血吸收光谱信号的信噪比,简化模型结构,提高模型的预测精度和计算效率,对推动血红蛋白浓度检测技术的发展具有进步意义。     

A Reduction Technique for Natural Gas Transmission Network Optimization Problems

We address the problem of minimizing the fuel consumption incurred by compressor stations in steady-state natural gas transmission networks. In the practical world, these type of instances are very large, in terms of the number of decision variables and the number of constraints, and very complex due to the presence of non-linearity and non-convexity in both the set of feasible solutions and the objective function. In this paper we present a study of the properties of gas pipeline networks, and exploit them to develop a technique that can be used to reduce significantly problem dimension, without disrupting problem structure, making it more amenable to solution.     

Compression strategies for the chemometric analysis of mass spectrometry imaging data

Application of chemometric methods to mass spectrometry imaging (MSI) data faces a bottleneck concerning the vast size of the experimental data sets. This drawback is critical when considering high‐resolution mass spectrometry data, which provide several thousand points for each considered pixel. In this work, different approaches have been tested to reduce the size of the analyzed data with the aim to allow the subsequent application of typical chemometric methods for image analysis. The standard approach for MSI data compression consists in binning mass spectra for each pixel to reduce the number of m/z values. In this work, a method is proposed to handle the huge size of MSI data based on the adaptation of a liquid chromatography‐mass spectrometry data compression method by the detection of regions of interest. Results showed that both approaches achieved high compression rates, although the proposed regions of interest–based method attains this reduction requiring lower computational requirements and keeping utter spectral information. For instance, typical compression rate reached values higher than 90% without loss of information in images and spectra.