Rapid and accurate detection of pesticide residue levels can help to prevent the harm of pesticide residue. This study used visible/near-infrared (Vis-NIR) (376–1044 nm) and near-infrared (NIR) (915–1699 nm) hyperspectral imaging systems (HISs) to detect the level of pesticide residues. Three different varieties of grapes were sprayed with four levels of pesticides. Logistic regression (LR), support vector machine (SVM), random forest (RF), convolutional neural network (CNN), and residual neural network (ResNet) models were used to build classification models for pesticide residue levels. The saliency maps of CNN and ResNet were conducted to visualize the contribution of wavelengths. Overall, the results of NIR spectra performed better than those of Vis-NIR spectra. For Vis-NIR spectra, the best model was ResNet, with the accuracy of over 93%. For NIR spectra, LR was the best, with the accuracy of over 97%, but SVM, CNN, and ResNet also showed closed and fine results. The saliency map of CNN and ResNet presented similar and closed ranges of crucial wavelengths. Overall results indicated deep learning performed better than conventional machine learning. The study showed that the use of hyperspectral imaging technology combined with machine learning can effectively detect the level of pesticide residues in grapes.

Grape is a fruit rich in various vitamins, and grape quality is increasingly highly concerned with by consumers. Traditional quality inspection methods are time-consuming, laborious and destructive. Near-infrared spectroscopy (NIRS) and hyperspectral imaging (HSI) are rapid, non-destructive and accurate techniques for quality inspection and safety assessment of agricultural products, which have great potential in recent years. The review summarized the applications and achievements of NIRS and HSI for the quality inspection of grapes for the last ten years. The review introduces basic principles, signal mode, data acquisition, analysis and processing of NIRS and HSI data. Qualitative and quantitative analysis were involved and compared, respectively, based on spectral features, image features and fusion data. The advantages, disadvantages and development trends of NIRS and HSI techniques in grape quality and safety inspection are summarized and discussed. The successful application of NIRS and HSI in grape quality inspection shows that many fruit inspection tasks could be assisted with NIRS and HSI.

Firmness, soluble solid content (SSC) and titratable acidity (TA) are characteristic substances for evaluating the quality of cherry tomatoes. In this paper, a hyper spectral imaging (HSI) system using visible/near-infrared (Vis-NIR) and near-infrared (NIR) was proposed to detect the key qualities of cherry tomatoes. The effects of individual spectral information and fused spectral information in the detection of different qualities were compared for firmness, SSC and TA of cherry tomatoes. Data layer fusion combined with multiple machine learning methods including principal component regression (PCR), partial least squares regression (PLSR), support vector regression (SVR) and back propagation neural network (BP) is used for model training. The results show that for firmness, SSC and TA, the determination coefficient R² of the multi-quality prediction model established by Vis-NIR spectra is higher than that of NIR spectra. The R² of the best model obtained by SSC and TA fusion band is greater than 0.9, and that of the best model obtained by the firmness fusion band is greater than 0.85. It is better to use the spectral bands after information fusion for nondestructive quality detection of cherry tomatoes. This study shows that hyperspectral imaging technology can be used for the nondestructive detection of multiple qualities of cherry tomatoes, and the method based on the fusion of two spectra has a better prediction effect for the rapid detection of multiple qualities of cherry tomatoes compared with a single spectrum. This study can provide certain technical support for the rapid nondestructive detection of multiple qualities in other melons and fruits.

Narrow-leaved oleaster (Elaeagnus angustifolia) fruit is a kind of natural product used as food and traditional medicine. Narrow-leaved oleaster fruits from different geographical origins vary in chemical and physical properties and differ in their nutritional and commercial values. In this study, near-infrared hyperspectral imaging covering the spectral range of 874–1734 nm was used to identify the geographical origins of dry narrow-leaved oleaster fruits with machine learning methods. Average spectra of each single narrow-leaved oleaster fruit were extracted. Second derivative spectra were used to identify effective wavelengths. Partial least squares discriminant analysis (PLS-DA) and support vector machine (SVM) were used to build discriminant models for geographical origin identification using full spectra and effective wavelengths. In addition, deep convolutional neural network (CNN) models were built using full spectra and effective wavelengths. Good classification performances were obtained by these three models using full spectra and effective wavelengths, with classification accuracy of the calibration, validation, and prediction set all over 90%. Models using effective wavelengths obtained close results to models using full spectra. The performances of the PLS-DA, SVM, and CNN models were close. The overall results illustrated that near-infrared hyperspectral imaging coupled with machine learning could be used to trace geographical origins of dry narrow-leaved oleaster fruits.

Radix Glycyrrhizae is used as a functional food and traditional medicine. The geographical origin of Radix Glycyrrhizae is a determinant factor influencing the chemical and physical properties as well as its medicinal and health effects. The visible/near-infrared (Vis/NIR) (376–1044 nm) and near-infrared (NIR) hyperspectral imaging (915–1699 nm) were used to identify the geographical origin of Radix Glycyrrhizae. Convolutional neural network (CNN) and recurrent neural network (RNN) models in deep learning methods were built using extracted spectra, with logistic regression (LR) and support vector machine (SVM) models as comparisons. For both spectral ranges, the deep learning methods, LR and SVM all exhibited good results. The classification accuracy was over 90% for the calibration, validation, and prediction sets by the LR, CNN, and RNN models. Slight differences in classification performances existed between the two spectral ranges. Further, interpretation of the CNN model was conducted to identify the important wavelengths, and the wavelengths with high contribution rates that affected the discriminant analysis were consistent with the spectral differences. Thus, the overall results illustrate that hyperspectral imaging with deep learning methods can be used to identify the geographical origin of Radix Glycyrrhizae, which provides a new basis for related research.

Early detection and diagnosis of crop anomalies is crucial for enhancing crop yield and quality. Recently, the combination of machine learning and deep learning with hyperspectral images has significantly improved the efficiency of crop detection. However, acquiring a large amount of properly annotated hyperspectral data on stressed crops requires extensive biochemical experiments and specialized knowledge. This limitation poses a challenge to the construction of large-scale datasets for crop stress analysis. Meta-learning is a learning approach that is capable of learning to learn and can achieve high detection accuracy with limited training samples. In this paper, we introduce meta-learning to hyperspectral imaging and crop detection for the first time. In addition, we gathered 88 hyperspectral images of drought-stressed tomato plants and 68 images of freeze-stressed tomato plants. The data related to drought serve as the source domain, while the data related to frost damage serve as the target domain. Due to the difficulty of obtaining target domain data from real-world testing scenarios, only a limited amount of target domain data and source domain data are used for model training. The results indicated that meta-learning, with a minimum of eight target domain samples, achieved a detection accuracy of 69.57%, precision of 59.29%, recall of 66.32% and $F_1$-score of 62.61% for classifying the severity of frost stress, surpassing other methods with a target domain sample size of 20. Moreover, for determining whether the plants were under stress, meta-learning, with a minimum of four target domain samples, achieved a detection accuracy of 89.1%, precision of 89.72%, recall of 93.08% and $F_1$-score of 91.37% outperforming other methods at a target domain sample size of 20. The results show that meta-learning methods require significantly less data across different domains compared to other methods. The performance of meta-learning techniques thoroughly demonstrates the feasibility of rapidly detecting crop stress without the need for collecting a large amount of target stress data. This research alleviates the data annotation pressure for researchers and provides a foundation for detection personnel to anticipate and prevent potential large-scale stress damage to crops.