Search

This study proposed a rapid microscopic examination method for pork freshness evaluation by using the self-assembled hyperspectral microscopic imaging (HMI) system with the help of feature extraction algorithm and pattern recognition methods. Pork samples were stored for different days ranging from 0 to 5 days and the freshness of samples was divided into three levels which were determined by total volatile basic nitrogen (TVB-N) content. Meanwhile, hyperspectral microscopic images of samples were acquired by HMI system and processed by the following steps for the further analysis. Firstly, characteristic hyperspectral microscopic images were extracted by using principal component analysis (PCA) and then texture features were selected based on the gray level co-occurrence matrix (GLCM). Next, features data were reduced dimensionality by fisher discriminant analysis (FDA) for further building classification model. Finally, compared with linear discriminant analysis (LDA) model and support vector machine (SVM) model, good back propagation artificial neural network (BP-ANN) model obtained the best freshness classification with a 100 % accuracy rating based on the extracted data. The results confirm that the fabricated HMI system combined with multivariate algorithms has ability to evaluate the fresh degree of pork accurately in the microscopic level, which plays an important role in animal food quality control.

In order to rapidly and efficiently determine geographical origin and characterization categories in five varieties of black Goji berry (Lycium ruthenicum Murr.), near-infrared (NIR) spectroscopy and chemometrics were utilized for data acquisition. Using this data, synergy interval partial least squares (Si-PLS), linear discriminant analysis (LDA), K-nearest neighbors (KNN), back propagation artificial neural network (BP-ANN) and least-squares support vector machine (LS-SVM) regression were systematically evaluated and compared during model development. LS-SVM was initially performed to calibrate the discrimination model to identify the geographical origins and categories of the black Goji berry samples. Compared with other models, the recognition rate of LS-SVM was more than 98.18 %, which showed excellent generalization for identification results. Total anthocyanin content was closely related with the quality of black Goji berry. Synergy interval partial least squares (Si-PLS) was applied to develop the prediction model of total anthocyanin content. The model was optimized by a leave-one-out cross-validation, and model performance was evaluated by assessing the root mean square error of the prediction (RMSEP) and correlation coefficient (R t ) in the prediction set. Experimental results showed that the optimum results of the Si-PLS model were achieved as follows: RMSEP = 0.602 mg/g and R t = 0.899 in the prediction set. The overall results sufficiently demonstrate that spectroscopy coupled with the Si-PLS regression tool has the potential to successfully discriminate black Goji berry varieties.

In multi-component spectral analysis, informative variables selection is important to get satisfied performance. The present research intends to establish relationship between eggs freshness and NIR spectroscopy, and build a compact and robust calibration model. Graphically-oriented local multivariate calibration modeling procedures were used comparatively to select efficient spectral variables in comparison to the full-spectrum model. Three kinds of methods, which were spectral interval selection, effective coefficient variables selection and genetic algorithm, were used for variable selection. Successive projections algorithm (SPA) showed its superior ability in reducing the complexity of model building. A satisfactory result was achieved while only 8 variables were used. Meanwhile, the optimal performance was obtained with genetic algorithm synergy interval partial least- square (GA-siPLS) by using 9 PCs and 42 variables selected, which resulted in root mean square error value of prediction (RMSEP) value of 3.29. This work indicates that it is feasible to identify egg freshness using NIR spectroscopy combined with graphically-oriented local multivariate analysis, and using variables methods is important to reduce he complexity of model building with fewer spectral variables and improve performance of calibration model. Keywords-Variable selection; Multivariate calibration; Near infrared spectroscopy; Egg freshness

A new room temperature gas sensor was fabricated with pigment-sensitized TiO2 thin film as the sensing layer. Four natural pigments were extracted from spinach (Spinacia oleracea), red radish (Raphanus sativus L), winter jasmine (Jasminum nudiflorum), and black rice (Oryza sativa L. indica) by ethanol. Natural pigment-sensitized TiO2 sensor was prepared by immersing porous TiO2 films in an ethanol solution containing a natural pigment for 24h. The hybrid organic-inorganic formed films here were firstly exposed to atmospheres containing methylamine vapours with concentrations over the range 2-10 ppm at room temperature. The films sensitized by the pigments from black-rice showed an excellent gas-sensitivity to methylamine among the four natural pigments sensitized films due to the anthocyanins. The relative change resistance, S, of the films increased almost linearly with increasing concentrations of methylamine (r=0.931). At last, the black rice pigment sensitized TiO2 thin film was used to determine the biogenic amines generated by pork during storage. The developed films had good sensitivity to analogous gases such as putrscine, and cadaverine that will increase during storage.

In protected agriculture, deficiency of an essential element may drastically affect plant growth, appearance and most importantly yield. Information about nutrient deficiencies in crops grown in controlled environment is essential to optimize food productivity. In this study, near infrared reflectance spectroscopy (NIRS) analysis was used to identify nitrogen (N) deficiency coupled with pattern recognition methods in mini-cucumber plants grown under non-soil conditions. Leaves at the first three nodes of nitrogen deficient plants and control plant were used for NIRS data acquisition. K-nearest neighbors (KNN) and artificial neural network (ANN) were applied to build diagnostics models, respectively. Some parameters of the model were optimized by cross-validation. The performance of the KNN model and the ANN model based on NIRS data was compared. Experiment results showed that the ANN model was better than the KNN model. The optimal ANN model was achieved when principle component factors were equal to 5 and identification rate of the ANN model were 100% in both the training set and the prediction set. This study demonstrated that the NIRS coupled with ANN pattern recognition method can be successfully applied to the diagnostics of nitrogen deficiency in minicucumber plant grown under non-soil conditions. Key words : Deficiency, nitrogen, near infrared reflectance spectroscopy (NIRS), artificial neural network.

Tin oxide gas sensor array based devices were often cited in publications dealing with food products. However, during the process of using a tin oxide gas sensor array to analysis and identify different gas, the most important and difficult was how to get useful parameters from the sensors and how to optimize the parameters. Which can make the sensor array can identify the gas rapidly and accuracy, and there was not a comfortable method. For this reason we developed a device which satisfied the gas sensor array act with the gas from vinegar. The parameters of the sensor act with gas were picked up after getting the whole acting process data. In order to assure whether the feature parameter was optimum or not, in this paper a new method called “distinguish indexâ€Â(DI) has been proposed. Thus we can assure the feature parameter was useful in the later pattern recognition process. Principal component analysis (PCA) and artificial neural network (ANN) were used to combine the optimum feature parameters. Good separation among the gases with different vinegar is obtained using principal component analysis. The recognition probability of the ANN is 98 %. The new method can also be applied to other pattern recognition problems.