Pesticide residues in fruits and vegetables pose a serious threat to food safety. Traditional detection methods have defects such as complex operation, high cost, and long detection time. Therefore, it is of great significance to develop rapid, non-destructive, and efficient detection technologies and equipment. In recent years, the combination of spectroscopic techniques and imaging technologies with machine learning algorithms has developed rapidly, providing a new attempt to solve this problem. This review focuses on the research progress of the combination of spectroscopic techniques (near-infrared spectroscopy (NIRS), hyperspectral imaging technology (HSI), surface-enhanced Raman scattering (SERS), laser-induced breakdown spectroscopy (LIBS), and imaging techniques (visible light (VIS) imaging, NIRS imaging, HSI technology, terahertz imaging) with machine learning algorithms in the detection of pesticide residues in fruits and vegetables. It also explores the huge challenges faced by the application of spectroscopic and imaging technologies combined with machine learning algorithms in the intelligent perception of pesticide residues in fruits and vegetables: the performance of machine learning models requires further enhancement, the fusion of imaging and spectral data presents technical difficulties, and the commercialization of hardware devices remains underdeveloped. This review has proposed an innovative method that integrates spectral and image data, enhancing the accuracy of pesticide residue detection through the construction of interpretable machine learning algorithms, and providing support for the intelligent sensing and analysis of agricultural and food products.

Near-infrared spectroscopy (NIRS) presents great potential in the identification of food adulteration due to its advantages of nondestructive, simple, and easy to operate. In this paper, a method based on NIRS and chemometrics was proposed to predict the content of common buckwheat (Fagopyrum esculentum Moench) flour in Tartary buckwheat (Fagopyrum tataricum (L.) Gaertn) flour. Partial least squares regression (PLSR) and support vector regression (SVR) models were used to analyze the spectrum data of adulterated samples and predict the adulteration level. Various preprocessing methods, parameter-optimization methods, and competitive adaptive reweighted sampling (CARS) wavelength-selection methods were used to optimize the model prediction accuracy. The results of PLSR and SVR modeling for predicting of Tartary buckwheat adulteration content were satisfactory, and the correlation coefficients of the optimum identification models were above 0.99. In conclusion, the combinations of NIRS and chemometrics indicated excellent predictive performance and applicability to analyze the adulteration of common buckwheat flour in Tartary buckwheat flour. This work provides a promising method to identify the adulteration of Tartary buckwheat flour and results obtained can give theoretical and data support for adulteration identification of agro-products.

Near-infrared spectroscopy (NIRS) holds significant promise in detecting food adulteration due to its non-destructive, simple, and user-friendly properties. This study employed NIRS in conjunction with chemometrics to estimate the content of low-price rice flours (Nanjing, Songjing, Jiangxi silk, Yunhui) blended with high-price rice (Wuchang and Thai fragrant). Partial least squares regression (PLSR), support vector regression (SVR), and back-propagation neural network (BPNN) models were deployed to analyze the spectral data of adulterated samples and assess the degree of contamination. Various preprocessing techniques, parameter optimization strategies, and wavelength selection methods were employed to enhance model accuracy. With correlation coefficients exceeding 87%, the BPNN models exhibited high accuracy in estimating adulteration levels in high-price rice. The SPXY-SG-BPNN, SPXY-MMN-BPNN, KS-SNV-BPNN, and SPXY-SG-BPNN models showcased exceptional performance in discerning mixed Wuchang japonica, Thai fragrant indica, and Thai fragrant Yunhui rice. As shown above, NIRS demonstrated its potential as a rapid, non-destructive method for detecting low-price rice in premium rice blends. Future studies should be performed to concentrate on enhancing the models’ versatility and practical applicability.

Tartary buckwheat has attracted more attention than common buckwheat due to its unique chemical composition and higher efficacy in the prevention of various diseases. The content of flavonoids in Tartary buckwheat (Fagopyrum tataricum (L.) Gaertn) is higher than that in common buckwheat (Fagopyrum esculentum Moench). However, the processing process of Tartary buckwheat is complex, and the cost is high, which leads to the frequent phenomenon of common buckwheat counterfeiting and adulteration in Tartary buckwheat, which seriously damages the interests of consumers and disrupts the market order. In order to explore a new and simple identification method for Tartary buckwheat and common buckwheat, this article uses metabolomics technology based on GC-MS to identify Tartary buckwheat and common buckwheat. The results show that the PLS-DA model can identify Tartary buckwheat and common buckwheat, as well as Tartary buckwheat from different regions, without an over-fitting phenomenon. It was also found that ascorbate and aldarate metabolism was the main differential metabolic pathway between Tartary buckwheat and common buckwheat, as well as the amino acids biosynthesis pathway. This study provides a new attempt for the identification of Tartary buckwheat and common buckwheat for the quality control of related agricultural products.