Application of hybrid image features for fast and non-invasive classification of raisin

A new approach for the non-invasive classification of raisins is presented based on the hybrid image features, namely morphological, color and texture features. A total of 74 features (8 morphological, 30 color, and 36 textural) were extracted from RBG images. Seven kinds of models were established based on different feature sets. They were three kinds of models established based on single feature set, three kinds of models established based on the combination of two feature sets, and one kind of model established based on the combination of all feature sets. Five kinds of classifiers, namely partial least squares (PLS), linear discriminant analysis (LDA), soft independent modeling of class analogy (SIMCA), and least squares support vector machine (LS-SVM) with linear and radial basis function (RBF) kernels were used for the model establishment based on different feature sets. The best correct answer rates (CAR) of 99% was obtained when LDA was used to establish the classification model based on the combination of all feature sets, which was higher than those of the models established based on single feature set or the combination of two feature sets. The results show that the feature combination is helpful to improve the accuracy of raisin classification. It was concluded that the varieties of raisin could be accurately classified based on RGB image features and the combination of morphological, color and texture features was an accurate way to improve the accuracy of classification.     

Image analysis of microstructural changes in almond cotyledon as a result of processing

Release of oil from nuts due to damaged cellular structures can degrade the quality of products incorporating nuts. The aim of this study was to investigate the effects of different processing conditions on microstructure of almond tissue and to quantify these changes using image processing. Spinning disk confocal fluorescence microscopy was used for imaging changes in microstructure of almonds as a function of different thermal processing of almonds. Multiple staining of Nile Red and Calcofluor White was applied to differentiate cell wall structures and oil bodies within individual almond cells without chemical fixation. An algorithm for image processing, included image preprocessing, segmentation, and determination of morphological features of segmented objects, was developed. Oil-roasting processes (140 °C and 150 °C) were found to have a significant impact on microstructure of almonds when compared to the hot air-roasting and blanching processes. Oil-roasted almond at 150 °C had a greater cellular damage due to cell wall and membrane rupture. These changes in microstructure of almonds would make them slightly more susceptible to release oil during storage. The image analysis presented allows quantitative evaluation for the effect of different processing on almond microstructure.     

CLASSIFICATION OF BARLEYS BASED ON MALTING QUALITY BY IMAGE ANALYSIS

Computerised image analysis was investigated as an automated quality control technique for rapid and precise classification of barleys according to their malting qualities. The nineteen samples used in this study were two-rowed winter type barleys. Six of these were assigned to Group 1 and the rest was included Into Group 2 according to their performance in a micro-malting system. Two sets of kernels from each barley sample were recorded, one for training and the other for testing purpose. To be able to have functions yielding the best discrimination scores in the training set, classification analysis was initially performed by the selection of each feature alone as an independent variable. Correct overall classification scores ranged between 52.9% and 78.1% when individual features were used in discriminant analyses. In the second step, the combination of features which yielded the best classification score was determined by stepwise discriminant analysis. Overall success in various combinations was more than 83%. To test the unknown barley samples, the discriminant function produced by the combination of compactness and equivalent diameter features was selected. Correct classification scores of this combination ranged between 60.0% and 94.7%. The result shows that the discrimination of the unknowns is very close to that achieved in the training set .     

A sandwich enzyme-linked immunosorbent assay for the detection of almonds in foods

An enzyme-linked immunosorbent assay was developed to detect almonds as potential allergenic contaminants in food. Polyclonal antibodies directed against roasted almonds were partially purified from immunized sheep and rabbits and used as capture and secondary antibodies, respectively, in a sandwich-type, 96-well plate format. Food samples and almond-spiked samples were extracted 1:10 in phosphate-buffered saline at 60 degrees C for 2 h, centrifuged, and applied to wells coated with sheep anti-almond antibody. After incubation, washing, and the addition of rabbit anti-almond antibody, the amount of almond present was detected with the subsequent addition of goat anti-rabbit immunoglobulin G-alkaline phosphatase conjugate and p-nitrophenyl phosphate substrate. Plate absorbances were read at 410 nm, and standard curves were developed in all matrices to quantify unknowns. Antibodies developed were specific for almond; however, some cross-reactivity was observed with extracts of some tree nuts and sesame seeds. Sodium dodecylsulfate-polyacrylamide gel electrophoresis and Western immunoblotting indicated that sheep anti-almond antibody recognized proteins extracted from black walnuts, Brazil nuts, cashews, hazelnuts, macadamia nuts, pistachios, and sesame seeds in addition to those from almond. The assay was optimized to detect less than 1 ppm of almond and was used successfully to determine almond residues in cereal and chocolate without cross-reacting interferences. A retail survey of 20 brands of cereal demonstrated that the assay produced statistically consistent results. This assay provides a useful quality control tool for the food industry for the protection of consumers allergic to almonds.     

Nutritional composition and protein value of the baru (Dipteryx alata Vog.) almond from the Brazilian Savanna

BACKGROUND: The almond of the baru tree (Dipteryx alata Vog.), a native species of the Brazilian Savanna, is used in the gastronomy of the central western region of the country. There is relatively little information about the chemical composition and nutritional value of the baru almond, which was the motivation for this research. RESULTS: The baru almonds had high lipid (397–437 g kg^?1) and protein (238–281 g kg^?1) contents. There were differences in the amino acid score (AAS = 83–103%) and limiting amount of sulfur amino acids, depending on the origin of the almond. The protein value of the baru almond was higher than that of the peanut according to the relative net protein ratio (RNPR_Baru = 74%, RNPR_Peanut = 66%) and the protein digestibility‐corrected amino acid score (PDCAAS). The baru almond also had high iron (mean 48.1 mg kg^?1), zinc (mean 46.6 mg kg^?1) and dietary fibre (mean 115.8 g kg^?1) contents in relation to Dietary Reference Intakes. CONCLUSION: The baru almond has a high nutrient density and high content of quality protein. Furthermore, the lipid and protein contents and amino acid profile of the baru almond are representative of edible seeds and similar to those of true nuts. This almond can be used as a complementary source of protein and as an excellent option for a healthy diet. Copyright © 2010 Society of Chemical Industry     

振动胁迫下双孢蘑菇高光谱成像品质检测

为快速有效识别双孢蘑菇轻微损伤,以不同振动时间后不同损伤程度的双孢蘑菇为研究对象,采集400～1 000 nm的完好无损、振动60 s和振动120 s双孢蘑菇的近红外高光谱图像,发现3 种类型的双孢蘑菇在450～750 nm的光谱曲线有明显差异。比较标准正态变量变换、SG（Savitzky-Golay）平滑和多元散射校正等预处理方法,确定SG平滑为最优预处理方法。并将处理后的数据采用连续投影算法和竞争性自适应重加权算法提取不同损伤程度的特征波段;基于灰度共生矩阵提取500 nm波长特征图像感兴趣区域的纹理特征,分别将光谱信息和纹理特征信息作为输入,建立偏最小二乘判别分析（partial least squares-discriminant analysis,PLS-DA）、BP（back propagation）神经网络和极限学习机损伤程度识别模型。结果表明,两种特征集建模,PLS-DA模型均表现出最好的识别效果,PLS-DA模型训练集和测试集平均识别准确率为93.33%、91.11%和88.89%、86.67%。最后基于光谱-纹理融合信息建立PLS-DA模型,训练集和测试集总体识别正确率分别为97.78%、95.56%。结果表明,光谱-纹理融合信息建模预测效果优于单一特征信息建立的判别模型。因此,采用高光谱融合信息建模可以提高不同损伤程度的双孢蘑菇检测精度,为双孢蘑菇贮藏、分类提供理论支撑。     

Comparison of volatile emissions from undamaged and mechanically damaged almonds

BACKGROUND: The navel orangeworm (NOW) Amyelois transitella (Walker) is a major insect pest of almonds causing considerable monetary setbacks for both growers and processors, and thus control of NOW is one of the top priorities for the almond industry. Field observations purport that NOW is attracted to previously injured almonds. Accordingly, in this study the volatile output of damaged almonds was investigated in an effort to identify potential attractants for further studies into the control and/or monitoring of NOW. Mature almonds from the Monterey variety were evaluated for their volatile composition after mechanical damage and compared with the volatile composition of undamaged almonds. RESULTS: Volatile organic compounds (VOCs) were collected on Tenax, desorbed and identified via gas chromatography/mass spectrometry analysis. VOCs unique to the damaged tree nuts included trace amounts of 3‐pentanol and isomers of the spiroketal chalcogran. VOCs that increased in relative amounts after damage include the spiroketal conophthorin and numerous four‐carbon ester and ketone as well as alcohol derivatives, in addition to two eight‐carbon chain compounds. CONCLUSION: Several VOCs, both unique and in increased amounts, were identified from damaged almonds. Their presence in damaged almonds warrants further investigation into their role in NOW response to damaged almonds, which may lead to insights into the control and/or monitoring of NOW. Published in 2008 by John Wiley & Sons, Ltd.     

A survey of the bacteriological quality of preroasted peanut, almond, cashew, hazelnut, and brazil nut kernels received into three Australian nut-processing facilities over a period of 3 years

There is little information about bacteriological quality of preroasted kernels available in the public domain. An investigation of the bacteriological quality of preroasted peanut, almond, cashew, hazelnut, and Brazil nut kernels received into three Australian nut-processing facilities was performed over a period of 3 years. A total of 836 samples were analyzed for aerobic plate count, and 921 samples for Salmonella and Escherichia coli. The 921 samples included 653 peanut, 100 cashew, 60 almond, 60 Brazil nut, and 48 hazelnut kernels. There was no E. coli detected in any sample. Salmonella subsp. II (Fremantle) was detected in one raw almond sample. The aerobic plate count percentages of positive samples with counts above the detection level of the plating method used (100 CFU/g) for peanuts, almonds, cashews, hazelnuts, and Brazil nuts were 84, 78, 74, 50, and 45%, respectively. Of the samples containing more than this detection limit, the means were 4.5, 4.4, 3.1, 2.5, and 3.8 log CFU/g respectively. Although roasted kernel quality was not within the scope of this survey, raw microbial bioload would be expected to reduce on roasting. The bacteriological quality of preroasted peanut, almond, cashew, hazelnut, and Brazil nut kernels received into nut-processing facilities in Australia does not appear to suggest a public health concern.     

Chromatographic Fingerprinting with Multivariate Data Analysis for Detection and Quantification of Apricot Kernel in Almond Powder

Adulteration of almond powder samples with apricot kernel was solved by gas chromatographic fatty acid fingerprinting combined with multivariate data analysis methods (principal component analysis (PCA), PCA-linear discriminant analysis (PCA-LDA), partial least squares (PLS), and LS support vector machine (LS-SVM). Different almond and apricot kernel samples were mixed at concentrations ranging from 10 to 90% w/w. PCA and PCA-LDA methods were applied for the classification of almonds, apricot kernels, and mixtures. PLS and LS-SVM were used for the quantification of adulteration ratios of almond. Models were developed using a training data set and evaluated using a validation data set. The root mean square error of prediction (RMSEP) and coefficient of determination (R ²) of validation data set obtained for PLS and LS-SVM were 5.01, 0.964 and 2.29, 0.995, respectively. The results showed that the methods can be applied as an effective and feasible method for testing almond adulteration.