基于太赫兹成像技术的发芽葵花籽无损检测研究

发芽葵花籽中脂肪和蛋白质等营养物质的含量会减少，影响油脂产品的产量及品质，因此需要保证葵花籽的品质安全。在对发芽葵花籽进行检测研究时，通常需要剥除葵花籽的外壳，破坏样本，这一操作不利于后续的贮藏加工。文章对发芽葵花籽的带壳无损检测方法进行了研究，采用太赫兹透射成像技术结合最大峰值(Max Peak Size)图像重构方法获取发芽葵花籽图像，应用自适应阈值分割法提取图像特征，应用连续投影(Successive Projections Algorithm,SPA)算法提取光谱特征，并通过串行融合实现特征融合，采用误差反向传播(Back propagation,BP)神经网络算法建立了发芽葵花籽在不同频率段内的识别模型。结果表明，采用SPA-BP算法建立的基于融合特征的发芽葵花籽识别模型效果较好，模型整体识别准确率为100%。由此可得，利用太赫兹时域透射成像技术可以实现发芽葵花籽的带壳无损检测，这对未来带壳油料作物品质的无损检测提供新思路。     

太赫兹波谱技术在木材工业中的应用

介绍了太赫兹波谱技术的基本特性,针对国外太赫兹波谱技术在木材工业中的应用研究现状,进行了分析。介绍了利用太赫兹时域光谱技术进行木材光学特性及含水率测定方法研究;介绍了利用太赫兹成像技术在木结构内部隐藏缺陷的无损检测等方向的研究成果。提出现阶段太赫兹波谱技术在木材工业领域的应用发展方向。     

基于太赫兹时域光谱技术的面粉品质快速无损检测研究

太赫兹(THz)波能够穿透大多数干的介电材料(塑料、陶瓷、衣物等),可实现对带包装物品的质量检测。为了研究THz光谱技术对带包装面粉的无损检测,首先对不带包装面粉进行太赫兹时域谱可行性分析及建模研究。对101份不同种类的面粉样本,用Tera Pulse 4000的太赫兹脉冲光谱仪采集了其太赫兹时域谱,对光谱预处理后,用偏最小二乘法(PLS)算法建立了面粉中3个指标水分、灰分、面筋的定量分析模型。各模型的预测相关系数都在0.89以上,研究结果表明,通过太赫兹时域光谱技术对面粉品质进行无损、快速检测具有可行性,对下一步太赫兹光谱技术直接对带包装的面粉进行检测研究奠定了坚实的基础。     

太赫兹衰减全反射技术对花生冻伤快速判别研究

目的 利用太赫兹衰减全反射光谱法(terahertz attenuated total reflection spectroscopy, THz-ATR)实现花生冻伤的快速鉴别。方法 实验选择种子公司购入的同品种冻伤和非冻伤花生各500粒,采集1000粒花生样本的0 359.97 THz光谱,通过光学参数计算得到样本集的吸光度、折射率和吸收系数。采用3点移动窗口平滑预处理和随机森林算法(Radom Forest)建立基于不同光学常数的花生冻伤识别模型。结果 在决策树棵数为500,特征变量数为38时,基于太赫兹吸光度建立的花生冻伤判别模型性能最佳,准确率、召回率、精确率达到97%、98%、96.1%。结论 本研究所建立的定性模型准确率高, THz-ATR技术有望为花生冻伤的快速无损鉴别提供一种新的高效的检测方法,为太赫兹技术在食品检测领域的应用提供了现实依据。     

基于高光谱成像技术与BO-XGBoost的花生含水率无损检测研究

含水率影响着花生的质量、储藏时长与出油率。本研究针对当前花生含水率测量效率低、有损检测、无法适应大规模检测等问题,探索基于高光谱成像技术的花生含水率无损快速检测方法。测量并建立了300份不同种类花生的高光谱原始图像及光谱数据集,并利用小波变换、多元散射校正(MSC)和一阶导数对数据进行预处理,结合PLS、XGBoost、BO-XGBoost算法建立花生含水量无损检测模型。通过实验对比得出,利用小波变换对原始光谱数据进行预处理后的光谱数据建立的BO-XGBoost模型最优,预测模型决定系数R²=0.953 9,均方根误差RMSE=0.806 5。实验表明,高光谱成像技术结合BO-XGBoost能够对花生含水率进行快速、准确、无损检测,且对其他农作物水分含量检测具有一定的借鉴意义。     

基于高光谱成像技术的苹果表面缺陷无损检测

以"红富士"苹果为研究对象,提出基于高光谱成像技术结合图像分割技术的苹果表面缺陷的无损检测方法。采用高光谱图像采集系统(400 nm～1 000 nm)采集完好无损和表面有缺陷苹果的高光谱图像;对采集到的高光谱图像进行最小噪声分离变换,提取感兴趣区域的平均光谱反射率;采用图像分割技术提出苹果表面缺陷的无损检测方法。结果表明:采用最小噪声分离变换可有效地消除苹果高光谱图像中的噪声;在700 nm～800 nm以及900 nm～1 000 nm波段范围内完好无损和表面有缺陷的苹果的光谱反射率值具有明显的差异,同时选取特征波长717.98 nm处的光谱反射率值小于0.6以及982.59 nm处的光谱反射率值大于0.52作为区分苹果正常区域和表面缺陷区域的阈值条件,进一步利用阈值分割方法对80个完好无损苹果和40个表面有缺陷苹果的正确识别率分别为97.5%和95%。表明高光谱成像技术结合图像分割技术可实现苹果表面缺陷的无损检测。     

基于太赫兹技术的小麦粉品质快速检测研究

该研究针对目前小麦粉品质方面检测方法存在的问题,提出利用太赫兹光谱技术对小麦粉进行快速无损品质检测研究。使用光谱仪与成像仪,采集了不同种类小麦粉样本的太赫兹光谱,使用TQ Analyst软件结合距离匹配法对小麦粉的太赫兹扫描光谱进行定性分析研究,富强粉和麦芯粉成功分类,模型性能指数达到88.9%,预测准确率达100%。使用OPUS软件结合偏最小二乘法（PLS）和一阶导数+矢量归一化（SNV）进行定量分析研究,水分定量模型R2为91.18%,交叉验证均方根为0.182;灰分定量模型 R2为83.37%,交叉验证均方根为0.064,最终通过实验结果分析得出太赫兹技术在食品品质检测方面的可行性。     

基于VIS/NIR高光谱成像技术检测鸡肉嫩度

利用400~1000 nm可见近红外高光谱成像系统对鸡肉嫩度进行快速无损检测研究。采集鸡肉表面的高光谱散射图像,提取样本感兴趣区域反射光谱曲线并用剪切力值表征鸡肉的标准嫩度。以原始光谱和多元散射校正(MSC)预处理光谱数据建立鸡肉嫩度的偏最小二乘回归(PLSR)模型,预处理光谱建立的模型效果更优。基于MSC预处理,采用偏PLS权重系数法结合逐步回归法筛选出了4个特征波长。然后采用PLSR和多元线性回归(MLR)模型分别建立特征波长处光谱反射值和鸡肉嫩度关系的数学模型,优选最佳模型。结果显示:MLR模型预测效果较好,预测相关系数(RP)和均方根误差(RMSEP)分别为0.94和1.97。研究表明:利用可见近红外高光谱成像技术结合多元回归分析法对鸡肉嫩度的快速无损检测是可行的。     

储粮品质太赫兹光谱检测技术研究进展

摘要：粮食质量与安全问题备受国民的关注,了解粮食在储藏过程中品质的变化趋势,特别是在早期阶段对其劣变状况进行快速、准确检测是当前粮食行业的重要任务之一。太赫兹光谱探测与成像技术具有快速、无损、衰减性小、无电离辐射伤害等特性现已成为无损检测技术的研究热点,在人体安全检查、环境监测、病变诊断、农产品质量控制等诸多领域取得了阶段性进展,在储粮品质检测方面也具有良好的应用前景。本文主要对太赫兹时域光谱技术的探测原理和光学参数提取以及成像技术进行了综述,重点阐述了该技术在储粮品质鉴别与分类、储粮新陈度、储粮真菌污染、以及储粮害虫检测方面的应用研究,并对太赫兹光谱技术在粮食品质快速检测中的发展趋势和应用前景进行了展望。     

基于高光谱成像及神经网络技术检测玉米含水率

基于高光谱成像及人工神经网络技术对玉米含水率进行了检测。检测波长为450～900nm,由玉米粒反射光谱图像获取反映其含水率的光谱特征波长。利用人工神经网络建立了玉米粒含水率的预测模型,模型相关系数达到0.98。对含水率预测结果的误差最大绝对值为2.1182,最小绝对值为0.0024。相对误差绝对值的平均值为0.3090,结果表明利用高光谱图像技术对玉米含水率进行无损检测是可行的。     

Influence of physical and biological variability and solution methods in fruit and vegetable quality nondestructive inspection by using imaging and near-infrared spectroscopy techniques: A review

Over the past decades, imaging and spectroscopy techniques have been rapidly developing and widely applied in nondestructive fruit and vegetable quality assessment. The physical properties (including size, shape, color, position, and temperature) and biological properties (including cultivar, season, maturity level and geographical origin) of fruits and vegetables vary from one to another. A great variety of physical and biological properties of agricultural products influence the optical propagation properties and interaction behaviors with incident light, thus decreasing the quality inspection accuracy. Many attempts have been made in image correction and spectral compensation methods to improve the inspection accuracy. This paper gives a detailed summary about influence of physical and biological variability, as well as the correction and compensation methods for eliminating or reducing the effects in fruit and vegetable quality nondestructive inspection by using imaging and spectroscopy techniques. The advantages and disadvantages of the solution methods are discussed and summarized. Additionally, the future challenges and potential trends are also reported.     

基于CNN神经网络的小麦不完善粒高光谱检测

利用高光谱成像技术对小麦不完善粒进行无损检测。以932个小麦为样本,其中正常粒样本486个、破损粒样本170个、虫蚀粒样本149个及黑胚粒样本127个为研究对象,通过高光谱图像采集系统采集样本的光谱信息,然后从每个样本的116个波段中选取30个波段,建立基于深度学习的卷积神经网络(convolutional neural networks,CNN)模型。实验中的CNN采用2个卷积层,第1层采用大小为3×3的32个卷积核,第2层采用大小为5×5的64个卷积核,池化层采用最大池,激活函数采用修正线性单元,为避免过拟合,在全连接层后面接入dropout层,参数设置为0.5,其他卷积参数均为默认值,得到校正集总识别率为100.00%,测试集总识别率为99.98%。最后,以支持向量机(support vector machine,SVM)为基线模型进行对比,从116个波段中选取90个波段进行建模,测试集总识别率为94.73%。通过实验对比可以看出,CNN模型比SVM模型识别率高。研究表明CNN模型能够实现对小麦不完善粒的准确、快速、无损检测。     

A Comprehensive Review on Food Applications of Terahertz Spectroscopy and Imaging

Food product safety is a public health concern. Most of the food safety analytical and detection methods are expensive, labor intensive, and time consuming. A safe, rapid, reliable, and nondestructive detection method is needed to assure consumers that food products are safe to consume. Terahertz (THz) radiation, which has properties of both microwave and infrared, can penetrate and interact with many commonly used materials. Owing to the technological developments in sources and detectors, THz spectroscopic imaging has transitioned from a laboratory‐scale technique into a versatile imaging tool with many practical applications. In recent years, THz imaging has been shown to have great potential as an emerging nondestructive tool for food inspection. THz spectroscopy provides qualitative and quantitative information about food samples. The main applications of THz in food industries include detection of moisture, foreign bodies, inspection, and quality control. Other applications of THz technology in the food industry include detection of harmful compounds, antibiotics, and microorganisms. THz spectroscopy is a great tool for characterization of carbohydrates, amino acids, fatty acids, and vitamins. Despite its potential applications, THz technology has some limitations, such as limited penetration, scattering effect, limited sensitivity, and low limit of detection. THz technology is still expensive, and there is no available THz database library for food compounds. The scanning speed needs to be improved in the future generations of THz systems. Although many technological aspects need to be improved, THz technology has already been established in the food industry as a powerful tool with great detection and quantification ability. This paper reviews various applications of THz spectroscopy and imaging in the food industry.     

近红外光谱波段优化在东北松子蛋白质定量检测中的应用

为了探究一种快速、无损与简便的东北松子品质检测方法,近红外光谱技术被应用到东北松子蛋白质无损检测研究中。利用偏最小二乘法建立带壳松子和去壳松仁的蛋白质定量分析模型,采用求导、多元散射校正、变量标准化校正、矢量归一化预处理方法优化模型,利用反向间隔偏最小二乘法、无信息变量消除法选取特征波段,建立全波段和特征波段下的偏最小二乘蛋白质预测模型。结果表明,带壳松子光谱经矢量归一化预处理方法后构建的模型最优,松仁光谱经变量标准化校正预处理方法后构建的模型最优;波段筛选能够优化模型质量,其中反向间隔偏最小二乘法的筛选结果最优,其带壳松子和松仁蛋白质模型校正集相关系数分别为0.9056和0.9383,验证集均方根误差分别为0.6670和0.5761。由此可知,经过优化后,模型的预测性能得到了提高,为带壳松子和松仁的蛋白质在线检测提供了一定的参考价值。     

Color image based sorter for separating red and white wheat

A simple imaging system was developed to inspect and sort wheat samples and other grains at moderate feed-rates (30 kernels/s or 3.5 kg wheat/h). A single camera captured color images of three sides of each kernel by using mirrors, and the images were processed using a personal computer (PC). Real time image acquisition and processing was enabled on an ordinary PC under Windows XP operating system using the IEEE 1394 data transfer protocol, DirectX application software, and dual-core computer processor. Image acquisition and transfer to the PC required approximately 17 ms per kernel, and an additional 1.5 ms was required for image processing. After classification, the computer could output a signal from the parallel port to activate an air valve to divert (sort) kernels into a secondary container. Hard red and hard white wheat kernels were used in this study to test and demonstrate sorter capability. Simple image statistics and histograms were used as features. Discriminant analysis was performed with one, two, or three features to demonstrate classification improvements with increased numbers of features. The sorter was able to separate hard red kernels from hard white kernels with 95 to 99% accuracy, depending on the wheat varieties, feed-rate, and number of classification features. The system is an economical and useful instrument for sorting wheat and other grains with high accuracy.     

The occurrence of aflatoxins in peanuts imported into Czechoslovakia for human consumption

Results of 492 analyses for aflatoxin in raw shelled peanuts imported into Czechoslovakia during 1982-1984 are presented. Most samples (55.3%) had aflatoxin content less than the detection limit of the radioimmunochemical screening method (0.8 micrograms/kg). Further analyses showed that 239 out of 410 samples of roasted peanuts contained aflatoxin below the detection limit. Only 1.9% of all peanut samples were found to have contamination level more than 5 micrograms/kg aflatoxin. The highest levels of aflatoxin observed were in a raw peanut sample containing 202.1 micrograms/kg and in a roasted peanut sample containing 32.6 micrograms/kg.     

Challenges and solutions of optical-based nondestructive quality inspection for robotic fruit and vegetable grading systems: A technical review

BackgroundOptical techniques, including computer vision, spectral imaging, near-infrared technology and other emerging imaging and spectroscopy techniques, have been rapidly developing and widely applied in fruit and vegetable grading systems for nondestructive quality inspecting and grading over the past decades. However, automatic detection of quality and grading is still difficult due to some still existing challenges, which are the key of blocking their commercialization in robotic fruit and vegetable grading systems. The challenges include the following aspects: the influence of physical and biological variability, whole surface detection, discrimination between defects and stems/calyxes, unobvious defect detection, robustness of the features and algorithms, as well as rapid optical detection system development. These challenges can reduce the fruit or vegetable quality inspection accuracy, thus greatly reducing automatic level of the quality inspecting and grading machines.Scope and approachAs agricultural engineers with about eight years of technical experience in fruit grading systems, we believe the ultimate goal of each scientific research should seek its task in serving the engineering. So, we have made many attempts to solve the challenges and increase the automation of the grading machines.Key findings and conclusionsThe review gives a detailed summary about the challenges and solutions of optical-based nondestructive quality inspection for fruit or vegetable grading systems from the perspective of engineering. Particular attention has been paid to the techniques that can improve the automation degree of the grading robot in this review. The advantages and disadvantages of the solutions are compared and discussed. Additionally, the remaining engineering challenges and future trends are also discussed.     

基于数字光处理技术的梨可溶性固形物含量的无损速测研究

目的为降低近红外光谱仪器制造成本,将近红外技术推广到农业生产一线,检验自主集成水果品质无损快速分析仪实验样机性能。方法以北京大兴产黄金梨、园黄梨为例,利用基于数字光处理技术内核的实验样机采集数据,采用偏最小二乘回归结合全交互验证算法分别建立黄金梨、园黄梨以及两种梨的可溶性固形物含量定量校正模型,并采用外部验证集对模型预测性能做进一步验证。结果黄金梨、园黄梨以及两种梨的可溶性固形物含量模型的测定系数R~2分别为0.6136、0.6576、0.5105,RMSEC分别为0.71、0.79、0.87:交互验证测定系数R~2_(CV)分别为0.5332、0.5076、0.4193,RMSECV分别为0.78、0.96、0.95;外部验证集相关系数r分别为0.7239、0.6825、0.6550,RMSEP分别为0.83、1.03、0.94。结论基于数字光处理技术内核自主集成的水果品质无损快速分析仪器在梨可溶性固形物含量的无损速测以及降低仪器制造成本方面具有一定的应用潜力。     

PREDICTION OF APPLE FRUIT FIRMNESS BY NEAR‐INFRARED MULTISPECTRAL SCATTERING1

The objective of this research was to predict fruit firmness by developing and evaluating a multispectral imaging system for real time acquisition of scattering images from apple fruit. A circular broadband light beam was used to generate light backscattering at the surface of apple fruit and scattering images were acquired, using a common aperture multispectral imaging system, from Red Delicious apple fruit for wavelengths at 680, 880, 905, and 940 nm. Scattering images were reduced to produce one‐dimensional spectral scattering profiles by radial averaging, which were then input into a backpropagation neural network for predicting apple fruit firmness. The neural network performed best when 10 neurons and 20 epochs were used. With three ratios of spectral profiles involving all four wavelengths, the neural network gave firmness predictions with the correlation of 0.76 and the standard error of 6.2 N for the validation samples.     

基于机器视觉技术的红枣全表面信息无损分拣系统研究与实现

设计了基于机器视觉技术相配套的红枣全表面信息无损分拣系统,整机由红枣自动喂料机构、红枣排序机构、输送及调整机构、分级执行机构、检测及控制部分组成。该系统利用镜面反射原理设计的红枣表面信息采集机构,能同时采集红枣正面、两侧面和底部图像,完成红枣全表面信息的获取,解决了以往分拣系统无法在高速运输状态下在线呈现红枣全表面信息的难点。基于视觉的无损分级软件将实时处理采集到的图像,分析与获取红枣的果形大小、表面质量、纹理等表面信息,并采用信息融合技术进行判别分级。试验结果表明,本系统分级速度快,分级准确率达到90%以上,可较好满足红枣自动化快速检测分级生产要求。