基于计算机视觉和机器学习的真伪卷烟包装鉴别

为解决人工鉴别真伪卷烟效率低、主观性强等问题,基于计算机视觉和机器学习建立了一种真伪卷烟包装鉴别模型。利用计算机视觉对卷烟包装进行图像处理和特征向量提取,分别以相似性度量模型、机器学习模型对特征向量进行分类并判定卷烟真伪。相似性度量模型采用曼哈顿距离模型进行分类,并对高斯双边滤波函数进行了参数优化;机器学习模型则以图像分块为基础,确定最优分块数量和面积。以"中华(软)""玉溪(软)""钻石(荷花)"3个卷烟品牌共603个真伪样品为对象,分别采用两种模型进行判定,结果表明:相似性度量模型在"玉溪(软)"样品测试集的准确率为96.17%;机器学习模型在"中华(软)""玉溪(软)""钻石(荷花)"3个样品测试集的准确率分别为98.99%、96.61%和100%。机器学习模型与相似性度量模型相比较,具有较好的迁移能力和鲁棒性,适用于卷烟真伪鉴别样品量大、品类多、图像复杂等情况。该方法可为提高真伪卷烟鉴别效率和准确率提供技术支持。     

基于烟用材料的衰减全反射红外光谱无损鉴别真假卷烟

为建立一种快速、无损、简便鉴别卷烟真伪的方法,采用衰减全反射红外光谱法(ATR-FTIR),分别采集多个品牌真品、假冒卷烟样品烟用材料相关部位的衰减全反射红外光谱,研究探索有鉴别力的鉴别点(即检测微区);通过直观比较及相似度(相关系数和QC比较)分析,选出多个分辨力强的鉴别点;采用Bootstrap法确定相似度判别阈值,建立了多个品牌卷烟真伪鉴别模型,并应用于未知样品的鉴别检验。结果表明:鉴别模型对于模型卷烟待测样品和非模型卷烟未知样品均能实现准确鉴别,鉴别正确率为100%;QC比较法具有较好的鉴别力,真品卷烟之间的QC值均大于94.620 7%,而真品与假冒卷烟(或非模型卷烟样品)的最大QC值为89.152 5%,低于90%;与现行感观鉴别法相比,该方法提升了卷烟真伪鉴别检验的科学性和准确性,可为烟草专卖打假提供另一种科学客观的方法。     

近红外技术在卷烟真伪鉴别中的应用

为了快速准确地鉴别卷烟的真伪,对A牌卷烟进行了近红外光谱扫描,采用光谱因子分析法建立了A牌卷烟的鉴别模型,并应用该模型对A牌(样品数88个)、非A牌(样品数80个)和假冒A牌卷烟(样品数10个)进行了鉴别。结果显示,A牌、非A牌和假冒A牌卷烟的鉴别准确率分别为92.0%、96.3%和100.0%,平均94.4%。该模型鉴别的准确性较高,可以作为卷烟真伪鉴别的一种辅助手段。     

基于t假设检验及SVM神经网络的卷烟真伪判定

为了提高物理指标鉴定卷烟真伪的可靠性,基于t假设检验和SVM神经网络,提出了一种利用卷烟物理指标鉴定卷烟真伪的模式识别模型。以卷烟库中随机选取的某品牌卷烟为例,采用该模型进行了卷烟真伪判定。判定结果显示:通过基于总通风率、滤嘴通风率及硬度3个物理指标的SVM神经网络分类预测模型,对示例卷烟的44个真假烟样本进行鉴别,正确率约为95%。说明基于t假设检验及SVM神经网络的模式分类方法可以进行卷烟真伪鉴别。     

基于近中红外光谱分析的大黄鱼产地区分技术

为实现4种产地来源的大黄鱼(福建养殖大黄鱼、温州养殖大黄鱼、舟山养殖大黄鱼和舟山野生大黄鱼)的准确快速区分,本研究应用傅里叶变换红外光谱仪对4种大黄鱼样本在4000～650 cm-1^的红外吸收指纹图谱进行测定,基于特征波段下的光谱吸收差异并结合PCA、CA和LDA、SVM模型对大黄鱼样品进行产地区分。结果如下：采用测定波段4000～650 cm-1^的全光谱采集信息经过SG预处理后建立的SVM模型效果最优,对4种大黄鱼样本的测试集准确率为83.3%。进一步对福建养殖大黄鱼和野生大黄鱼的产地区分方法优化后,选取特征波段3690～2800 cm-1⁺¹⁸⁰⁰～650 cm-1^{的光谱信息}经过1st ^Der、2nd ^Der和SNV 3种方式预处理后建立LDA判别模型,光谱训练集与预测集的准确率均达到100%;3690～2800 cm-1⁺¹⁸⁰⁰～650 cm-1^{的光谱信息}经SNV、MSC预处理后的PCA效果最佳,2种大黄鱼样本间彼此间距远、无重叠,且前两个主成分累计贡献率均在90%以上;经SNV预处理后的CA分析中除21号野生大黄鱼被误聚到福建养殖大黄鱼类中,其余2种产地大黄鱼样本各自聚为一类。研究表明基于近中红外光谱测定并结合化学计量学处理的方法能够对大黄鱼产地进行较准确地快速区分,从而为大黄鱼溯源鉴别、维护消费者权益和保护区域大黄鱼产业等方面提供技术支撑。     

基于可见光谱的颜料组分鉴别研究

利用多光谱技术对彩绘艺术品颜料成分及密度进行鉴定,是当前文化领域面临的重要课题。本研究首先通过对颜料可见光反射光谱特征波段进行分析,将判别能力最强的特征波段引入鉴别模型,而后,基于判别模型实现颜料组分及含量的鉴别。实验数据表明,本研究采用的方法,对三色油墨混合组分的鉴别总体效果比较理想,其中青色油墨与品红油墨组分鉴别的准确率达到100%,黄色油墨组分的鉴别总体准确率为94.1%;结合多光谱图像获取技术,本研究提出的方法能够实现彩绘文物及实验品的颜料组分映射,对艺术品复制及修复具有重要意义。     

湖北金三峡:顺应趋势 寻求突围

<正>成功的烟包企业并不一定就能成为成功的社会包装企业,首先要意识到行业间的差异。湖北金三峡目前主要生产红金龙、红塔山、红河、云烟等品牌的卷烟包装,年产量为100万大箱,产值规模在6亿元左右。目前烟包行业的利润空间确实在缩小,烟草行业卷烟品牌的集中整合,也带来配套烟包行业的重新洗牌。行业的整合对我们来说既是挑战,也是机遇。如何在留住原有客户、原有品牌的同时,开发新客户、新品牌,是企业发展的重中之重。烟包企业唯有苦练内功,才能顺利突围。     

基于二维相关近红外光谱的白酒酒龄鉴别

目的:杜绝以低龄酒冒充高龄酒,实现对白酒酒龄的判别。方法:利用近红外光谱仪采集白酒样品的近红外透射光谱,以酒龄作为外扰,构建各白酒样品的同步和异步二维相关谱;在此基础上,提取每个样品的自相关谱,并结合马氏距离法建立白酒酒龄判别模型。结果:1400~1800 nm波段范围内包含白酒样品随酒龄变化的有效信息;模型校正集判别正确率为93.3%,预测集判别正确率为92.0%。结论:二维相关谱结合马氏距离法能够对白酒酒龄进行有效鉴别。     

基于高光谱成像的羊肉掺假可视化无损定量检测

目的：快速、准确检测羊肉掺假。方法：利用可见—近红外(400~1 000 nm)和短波近红外(900~1 700 nm)高光谱成像仪对羊肉中掺假不同比例的鸭肉进行数据采集,比较两个波段范围内不同光谱预处理方法的偏最小二乘法(PLS)建模效果,最终在可见—近红外波段选择归一化预处理方法,在短波近红外波段选择标准正态变量变换(SNV)预处理方法。分别对两个波段的光谱数据进行最优的预处理后,采用连续投影算法(SPA)、竞争性自适应重加权算法(CARS)、区间随机蛙跳算法(iRF)和组合区间偏最小二乘法(SiPLS)对特征波长进行选取。结果：在短波近红外(900~1 700 nm)波段采用SNV-SPA-PLS模型的羊肉掺假预测效果最好,预测集决定系数为0.968 4,预测标准偏差为0.058 2,预测集相对分析误差为5.625 4,并得到较好的图像反演结果。结论：利用不同波段的高光谱成像技术可实现对羊肉掺假的快速无损定量检测。     

基于深度学习卷积神经网络的花生籽粒完整性检测

目的:准确区分完整花生、果仁破损花生和表皮破损花生。方法:提出了一种基于深度学习卷积神经网络(CNN)的花生籽粒完整性检测方案。搭建了花生籽粒色选系统,建立了花生籽粒图像库;利用改进的密度峰值聚类(DPC)算法对CNN卷积核进行自适应压缩,有效平衡网络深度和运算效率;采用改进的麻雀搜索算法对CNN超参数配置和网络结构进行优化,得到适用于花生籽粒完整性检测的CNN模型。结果:相比于DL-CNN、CO-Net等检测方法,该方案识别准确率提高了5.41%~13.92%,花生籽粒单幅图像检测时间缩短了约16.9%。结论:该方法可有效提高花生籽粒完整性检测的准确率和实时性。     

卷烟主流烟气气溶胶分布研究

对卷烟主流烟气气溶胶分布进行了研究,按照标准吸烟条件,对4个卷烟进行了测试,通过在线200倍稀释,采用测量范围在7～10000 nm的电子低压撞击器对气溶胶粒径和浓度进行了测试。试验结果表明,不同卷烟主流烟气气溶胶分布有一定的差异,烟气气溶胶浓度随抽吸口数增加而增加,所测试的A、D样品气溶胶最高浓度粒径出现在0.261～0.381μm,B、C样品卷烟气溶胶最高浓度粒径出现在0.028μm;A、D样品卷烟气溶胶平均粒径为0.23μm,而B、C样品卷烟平均粒径为0.16μm;4种卷烟主流烟气气溶胶平均浓度介于1.09&#215;109～2.21&#215;109个/cm3之间,而相应每支烟的总粒子数介于2.29&#215;1011～3.88&#215;1011个粒子。     

Concentrations of migrated mineral oil/polyolefin oligomeric saturated hydrocarbons (MOSH/POSH) in Chinese commercial milk powder products

ABSTRACT

Mineral oil contaminants migrated from food contact materials have raised much concern in the past few decades. However, survey data of the occurrence of mineral oils in foods are only available for a limited variety of foods, which do not include the contaminants in Chinese milk powders. Thus, to conduct an analysis of mineral oils migrated into milk powder products from different packaging materials, 50 Chinese commercial samples (including 38 infant formulas), which were packaged in metal cans, paper containers, paper boxes and aluminium foil-plastic bags, were analysed. Mineral oil saturated hydrocarbons (MOSH), polyolefin oligomeric saturated hydrocarbons (POSH) and aromatic hydrocarbons (MOAH) migrated into those samples were extracted and quantitatively analysed by on-line liquid chromatography coupled with gas chromatography (LC-GC). The results indicated no surface MOAH was detected in any of the samples, while 33 samples contained MOSH/POSH with surface contents of 0.10 ~ 5.09 mg kg^?1. Moreover, the amounts of those MOSH/POSH are closely related to packaging materials, among which the surface contamination values of products in metal cans were the lowest, followed by products in paper containers and boxes, and the values of samples in aluminium foil-plastic bags were the highest. In addition, the surface MOSH/POSH content exhibited a positive correlation with fat content in the samples due to their similar polarities.     

YB517盒外透明纸包装机烟包侧封整形装置设计与试验

目的:解决YB517盒外透明纸包装机烟包侧封面透明纸皱褶、不平整等质量问题。方法:设计了一种烟包侧封整形装置。采用YB517盒外透明纸包装机输出转盘接烟板驱动轴作为动力源,通过设计往复运动机构、加热器组件、烟包定位组件以及优化推包座等部件,实现烟包侧封的加热整形;并运用Solidworks软件对整形装置的位移、速度、加速度进行仿真分析。接入“YB517拉带热封器”的温度检测控制系统接口,在IPC控制面板上实现温度参数的显示和调节。采用上机试验取样抽检的方法对改造前后盒装透明纸折皱情况进行对比分析。结果:该侧封整形装置运动无干涉,速度、加速度曲线平缓,无明显冲击,运行稳定。YB517盒外透明纸包装机安装侧封整形装置后,盒装透明纸折皱比例由改造前的11.58%降低到2.50%,下降78.40%。结论:该侧封整形装置能够减少皱褶,保证产品质量。     

PRELIMINARY EVALUATION OF THE APPLICATION OF THE FTIR SPECTROSCOPY TO CONTROL THE GEOGRAPHIC ORIGIN AND QUALITY OF VIRGIN OLIVE OILS

ABSTRACT

A rapid Fourier transform infrared (FTIR) attenuated total reflectance spectroscopic method was applied to determine qualitative parameters such as free fatty acid (FFA) content and the peroxide value (POV) in virgin olive oils. Calibration models were constructed using partial least squares regression on a large number of virgin olive oil samples. The best results (R² = 0.955, root mean square error in cross validation [RMSECV] = 0.15) to evaluate FFA content expressed in oleic acid % (w/w) were obtained considering a calibration range from 0.2 to 9.2% of FFA relative to 190 samples. For POV determination, the result obtained, built on 80 olive oil samples with a calibration range from 11.1 to 49.7 meq O₂/kg of oil, was not satisfactory (R² = 0.855, RMSECV = 3.96). We also investigated the capability of FTIR spectroscopy, in combination with multivariate analysis, to distinguish virgin olive oils based on geographic origin. The spectra of 84 monovarietal virgin olive oil samples from eight Italian regions were collected and elaborated by principal component analysis (PCA), considering the fingerprint region. The results were satisfactory and could successfully discriminate the majority of samples coming from the Emilia Romagna, Sardinian and Sicilian regions. Moreover, the explained variance from this PCA was higher than 96%.

PRACTICAL APPLICATIONS

The verification of the declared origin or the determination of the origin of an unidentified virgin olive oil is a challenging problem. In this work, we have studied the applicability of Fourier transform infrared coupled with multivariate statistical analysis to discriminate the geographic origin of virgin olive oil samples from different Italian regions.

     

Fourier transform infrared spectroscopy (FTIR) as a tool for discriminating <Emphasis Type="Italic">Salmonella typhimurium</Emphasis> contaminated beef

Detection of beef contamination from harmful pathogens will be helpful in protecting the consumer safety and controlling the outbreaks. In this paper, the potential of Fourier transform infrared spectroscopy (FTIR) was investigated to discriminate the Salmonella contaminated packed beef. A suitable headspace sampling system was designed and used to collect the headspace volatiles from the packed meat to the FTIR gas cell. Spectral signatures of headspace volatiles of meat packages were used to classify the packed meat samples as contaminated or not. FTIR spectrum was divided into several regions in order to reduce the dimensionality as well as to select the regions based on the absorbance properties of various volatiles present in headspace of meat package. Principal component analysis was performed on the entire spectrum (4000–500 cm⁻¹) as well as on the selected sub-regions of entire spectrum. Two statistical classification techniques (linear and quadratic discriminate analysis) were used to develop classification models. The statistical models were validated using bootstrap cross validation technique. The total average classification accuracies were evaluated in terms of coefficient of variance (% CV). Based on the mean of total average classification accuracies and its % CV calculated from five similarly conducted experiments, it was found that the statistical models developed on a part of the spectra (500–850 cm⁻¹) and full spectra (4000–500 cm⁻¹) can be used as potential classification models for non-destructive discrimination of Salmonella contaminated packed beef samples from uncontaminated ones. These results need to be further validated on dataset with larger sample size.     

烟用爆珠内液质量稳定性检测——基于紫外光谱技术结合SVR算法

  目的  基于紫外（UV）光谱技术和支持向量机回归（SVR）算法,建立一种烟用爆珠内液判别和稳定性分析方法。  方法  采用无水乙醇稀释爆珠内液,UV扫描,光谱经预处理后,建立四种不同牌号烟用爆珠SVR模型,并对模型进行了验证。  结果  （1）最佳UV图谱预处理方法是平滑后归一化;最佳SVR类型是ν-SVR;最佳核函数是径向基核函数。（2）四个牌号烟用爆珠内液ν-SVR模型校正集分类变量的预测值与实测值的相关系数均≥0.9993,SVR模型对于参与建模的400个烟用爆珠内液样本（校正集）和未参与建模的80个烟用爆珠内液样本（验证集）的预测准确度均为100%,模型拟合性好,预测精度高,判别能力强。（3）基于SVR模型分类变量值建立的单值控制图可以对烟用爆珠内液的稳定性进行快速判定,判定结果与GC/MS检测结果一致。  结论  紫外可见光谱技术结合SVR算法可对不同牌号烟用爆珠内液质量进行有效判别,且方法快速准确、经济环保、易于推广。      

基于三角测量的卷烟包装密封度无损检测方法

目的：解决现有卷烟包装密封度检测过程中出现的耗时长、破坏样品等问题。方法：将卷烟包装放置于密封负压环境中，利用三角测量原理检测卷烟包装在负压作用下体积变化情况，结合流体分析，研究卷烟包装体积变化与密封度之间关系。结果：泄漏孔直径在0.2~0.8 mm时，泄露孔直径与卷烟包装最大离面位移的偏差以及泄漏孔直径与卷烟包装最大离面位移增长率保持特定值时间的偏差呈逐渐减小的趋势，根据泄漏孔直径的不同，能够清晰分辨出差距；在0.4~0.8 mm时，经过3~7 s的检测则能根据变形快速获取泄漏孔直径，且最大离面位移与泄漏孔直径之间相关系数超过0.9，存在强相关性。经测试验证，该方法检测出合格品的准确度为88.1%。     

卷烟燃烧锥形态参数测定及与燃烧速度间的关系

为深入了解卷烟的燃烧性能,建立了一种基于图像处理技术的卷烟燃烧锥长度、面积、偏离角等形态参数的测定方法,通过48种卷烟样品自由燃烧速度及燃烧锥形态参数的测定,研究了卷烟燃烧锥形态参数与自由燃烧速度间的关系。研究结果表明:（1）48种卷烟燃烧锥锥长、锥面积、偏离角的RSD分别介于2.7%~10.1%、1.1%~10.1%、12.0%~78.3%之间,其中燃烧锥长度及面积的重复性较好,依据测量结果区分样品间差异性的灵敏度较高。（2）48种卷烟的燃烧锥锥长、锥面积和偏离角分别介于6.604~14.017mm、38.729~92.825mm2、2.778~7.636°之间,不同牌号卷烟的三项燃烧锥形态参数均存在极显著差异性。（3）卷烟自由燃烧速度与燃烧锥锥长和面积间均存在极显著正相关关系,且两者之间存在交互作用。      

Plasticized methylcellulose coating for reducing oil uptake in potato chips

BACKGROUND: As a result of consumers' health concerns and the trend towards healthier and low‐fat food products, research has been undertaken to reduce the amount of fat absorbed in fried foods. This work focused on studying the efficacy of sorbitol and glycerol as plasticizers of methylcellulose coatings used to reduce oil uptake during the frying process of potato chips RESULTS: Changes in color, mechanical properties, water activity and lipid oxidation during storage were monitored. Also, an explanation regarding the different performances between both methylcellulose coatings with and without plasticizer was attained and techniques from the field of packaging films such as dynamic mechanical analyzer (DMA) and Fourier transform infrared spectroscopy were applied to analyze the behavior of coatings submitted to the frying operation. The application of a methylcellulose coating was an adequate choice to reduce oil absorption in fried potato chips. The most effective formulation was 10 g L^?1 methylcellulose with the addition of 7.5 g L^?1 sorbitol. With the incorporation of this formulation, oil absorption was reduced by 30%. Neither the sorbitol concentration nor the presence of the MC coating affected the puncture maximum force and color parameters L and a*. The results of the sensory analysis indicated that the panelists could not distinguish between the coated and uncoated potato chips. CONCLUSION: Methylcellulose‐based coating plasticized with sorbitol could be an alternative for obtaining healthier potato chips. Copyright © 2011 Society of Chemical Industry     

Rapid classification of virgin and recycled EPS containers by Fourier transform infrared spectroscopy and chemometrics

ABSTRACT

A rapid and sensitive method for classification of virgin and recycled expanded polystyrene (EPS) food containers was developed using Fourier transform infrared spectroscopy (FTIR) and chemometrics. This method includes preparing a transparent film by dissolution, examining by FTIR and developing classification models. The degradation of EPS containers occurring during the recycling process was reflected by the carbonyl region of the infrared spectrum which was used as variables for multivariate data analysis. PCA was used to reduce the data dimension and view the sample similarities. Soft independent modelling of class analogy (SIMCA), partial least squares-discrimination analysis (PLS-DA) and linear discrimination analysis (LDA) were applied to construct three classification models. The best discrimination results were obtained by an LDA model, with all samples correctly classified. PLS-DA and SIMCA could not classify the recycled EPS samples with low levels of adulteration. When applying this method to commercially available EPS containers, about 45% of samples were shown to contain recycled polystyrene resins. It is concluded that the carbonyl region of the infrared spectra coupled with chemometrics could be a powerful tool for the classification of virgin and recycled EPS food containers.