In order to distinguish different varieties of Lycium barbarum effectively, a fast nondestructive detection method based on hyperspectral imaging technology was proposed. Six varieties of L. barbarum were selected as the research objects to obtain hyperspectral images. With the whole L. barbarum taken as the object, the region of interest was obtained by threshold segmentation, and the average spectra value of the image points of a single L. barbarum was extracted as the spectral data of the sample. Initially, standard normalized variate was used to preprocess the original spectral data. Furthermore, compared with other methods, competitive adaptive reweighted sampling (CARS) was chosen to extract the characteristic wavelengths. Additionally, the model of support vector machine (SVM) was set. The results showed that the SVM model based on CARS had the best classification effect. The training set accuracy was 100%, and the prediction set accuracy was 85%. Finally, in order to improve the classification accuracy, the whale optimization algorithm (WOA) was introduced. The accuracy of training set and prediction set obtained by WOA‐SVM model were 89.44 and 88.33% respectively. Therefore, it was feasible to use hyperspectral imaging technology combined with CARS‐WOA‐SVM model to identify different varieties of L. barbarum.

In order to identify the freshness grade of eggs nondestructively and rapidly, hyperspectral imaging technology was used in this article. The hyperspectral data of 200 samples of three freshness grades was acquired by using hyperspectral image acquisition system (400.68–1,001.612 nm), and then the freshness grade of egg samples was measured by stoichiometry. First, Mahalanobis distance algorithm was used to remove abnormal sample data. Second, savitzky–golay and wavelet threshold denoising combined with standard normalized variable were used to pretreat the spectral data, respectively. Third, iteratively retains informative variables (IRIV), variable iterative space shrinkage approach, and competitive adaptive reweighted sampling were used for feature wavelength selection. Since the classification accuracy of support vector machine (SVM) model was affected by the selection of parameters, genetic algorithm (GA) was introduced to search the optimal parameters in SVM and compared with grid search algorithm. Finally, the result indicated that the classification accuracy of training set and test set of the optimal classification model (IRIV‐GA‐SVM) reached 99.29% and 97.87%, respectively. Thus, it is feasible to use hyperspectral image technology to detect egg freshness grade. Practical Applications The freshness grade is one of the most important indexes to measure the quality of eggs. Traditional methods of detecting egg freshness grade are time‐consuming and destructive, which cannot meet the test needs of modern agriculture. Hyperspectral imaging, an emerging technology, can provide both spectral and spatial information simultaneously, and has the advantages of nondestructive, fast and nonpollution. The result indicated that hyperspectral imaging technology for the detection of freshness grade of eggs is feasible.

In the process of tea production and storage, mold is constantly multiplying due to improper production methods and environmental conditions. To realize the nondestructive detection of the total mold colony count in green tea, an accurate and rapid method based on visible-near-infrared (431-962 nm) hyperspectral image was proposed. Firstly, the spectral data extracted from hyperspectral images was preprocessed and partial least squares regression model based on different preprocessing methods was established to determine the best preprocessing method. Then, competitive adaptive reweighted sampling (CARS) and variable combination population analysis were used to select the characteristic wavelengths and support vector regression (SVR) was introduced to establish quantitative detection model. Because the parameter setting of SVR directly affects the effect of the model, a combination of genetic algorithm (GA) and particle swarm optimization (PSO) was adopted to optimize the parameters c (penalty factor) and g (kernel function parameter). The results showed that based on the wavelength selected by CARS, the SVR model optimized by GA-PSO (CARS-GA-PSO-SVR) achieved accuracy with R 2 P of 0.9577 and root mean square error of prediction set of 0.1140 lg(CFU/g). Therefore, hyperspectral imaging technology can realize the nondestructive determination of the total mold colony count in green tea. Practical applications Some molds, such as aspergillus and penicillium, are the main factors of tea mildew. These molds can produce mycotoxins such as aflatoxin and citreoviridin, which can not only cause damage to the tea quality, but also threaten the health of tea drinkers. In this paper, the total mold colony count of green tea was studied based on hyperspectral imaging technology. The experiment result indicated that the use of hyperspectral imaging technology can achieve accurate, nondestructive, and rapid detection of the total mold colony count in green tea. This research provides an effective solution to the quantitative detection of the total mold colony count in green tea. 1 | INTRODUCTION Tea is one of the most popular beverages in the world, and it is rich in abundant organic chemical components, inorganic minerals, nutrients, and medicinal ingredients (Muhammad et al., 2017). In addition, tea Yan Cao and Haoran Li are co-first authors.

Accurate, rapid, and nondestructive identification of rice seed varieties has great significance for agriculture and food security, a method based on information fusion and artificial fish swarm algorithm (AFSA) combined with the hyperspectral imaging (HSI) of five kinds of rice seeds was proposed in this work. First, the spectral and image data were obtained from HSI, and the spectral data were preprocessed by detrending. Then, bootstrapping soft shrinkage (BOSS), variable iterative space shrinkage approach, successive projections algorithm, and principal component analysis were adopted to select feature variables from the spectral and image data. Next, the support vector machine (SVM) model was constructed based on the spectral and image feature variables. In order to further improve the classification accuracy of single feature model, the model based on fused feature was developed and it was finally optimized by AFSA. The research showed that the feature variables (114 spectral variables and 11 image variables) selected by BOSS were representative, and the model accuracy based on BOSS spectral and image features reached 91.48% and 70%, respectively. The performance of the SVM model based on fused feature was improved significantly, and the model accuracy reached 97.22%. After AFSA optimization, the model accuracy finally reached 99.44%. The above result confirmed that using AFSA to optimize the model based on fused feature could be a promising method to identify rice seed varieties. Practical application Rapid and accurate identification of rice seed varieties contributes to establishment of online rice seed identification system. HSI combined with information fusion and AFSA can overcome the disadvantage of low accuracy of traditional nondestructive testing methods. The method proposed in this article can be recommended to be widely popularized in farms, grain markets, and market regulators.

In this study, a rapid and non-destructive method for the classification of tea varieties based on fluorescence hyperspectral imaging technology was proposed in the wavelength range of 400.6797-1001.612 nm. Multiplication Scatter Correction (MSC) was used to preprocess the spectral data of tea samples. For optimal feature selection, variable iterative space shrinkage approach (VISSA) and competitive adaptive reweighed sampling (CARS) were established and CARS achieved good results on tea spectral data. Four linear and non-linear classification models, Naïve Bayes (NB), K-Nearest Neighbors (KNN), Support Vector Machine (SVM), and Artificial Bee Colony Support Vector Machine (ABC-SVM) were established and then performances of classification models were compared according to classification accuracy. The classification accuracy of the ABC-SVM model coupled with CARS was achieved 100% which was the highest classification accuracy. The results of this study demonstrated that fluorescence hyperspectral image technology combined with the CARS-ABC-SVM model is feasible to classify tea varieties. Novelty Impact Statement: Traditional methods for the classification of tea varieties mainly focus on the appearance of tea and depend on human sensory evaluation, which is expensive and time-consuming. In this study, a method involving fluorescence hyperspectral image technology with the CARS-ABC-SVM algorithm successfully was used for precise and non-destructive classification of tea varieties. 2 of 9 | AHMAD et Al. spectral information of an object by integrating conventional spectroscopic and imaging techniques (Kaliramesh et al., 2013;