Advances of Vis/NIRS and imaging techniques assisted by AI for tea processing

Wu P., Liu J., Jiang M., Zhang L., Ding S., Zhang K.

Song Y., Yi W., Liu Y., Zhang C., Wang Y., Ning J.

Yu Y., Li Q., Hua Z., Yin C., Shi Y.

Choudhury M., Elyoseph Z., Fast N.J., Ong D.C., Nsoesie E.O., Pavlick E.

Nature Reviews Psychology invited six researchers from cognitive science, clinical psychology, social psychology, language science and public health to share their perspectives on current and future uses of generative artificial intelligence, including its impacts on research and humankind.

Zhao S., Adade S.Y., Wang Z., Jiao T., Ouyang Q., Li H., Chen Q.

Artificial intelligence (AI) technology is advancing the digitization and intelligence development of the food industry. A promising application is using deep learning-assisted visible near-infrared (vis-NIR) spectroscopy to monitor residual sugar and bacterial concentration in real-time, ensuring kombucha quality during production. The feature fingerprints of residual sugar and bacterial concentration were extracted by four variable selection algorithms and then reconstructed using serial and parallel processing methods. Based on these reconstructed features, Partial Least Squares (PLS) and Convolutional Neural Networks (1DCNN and 2DCNN) models were developed and compared. The experimental results showed that the 2DCNN model based on reconstruction features achieved superior performance. The RPDs of the residual sugar and bacterial concentrations models were 4.49 and 6.88, while the MAEs were 0.42 mg/mL and 0.04 (Abs), respectively. These results suggest that the proposed modeling strategy effectively supports quality control during kombucha production and provides a new perspective for spectral analysis.

Zhu F., Wang J., Zhang Y., Shi J., He M., Zhao Z.

Huang M., Tang Y., Tan Z., Ren J., He Y., Huang H.

Xu Q., Zhou Y., Wu L.

Zheng P., Solomon Adade S.Y., Rong Y., Zhao S., Han Z., Gong Y., Chen X., Yu J., Huang C., Lin H.

During the fermentation process of Oolong tea, significant changes occur in both its external characteristics and its internal components. This study aims to determine the fermentation degree of Oolong tea using visible–near–infrared spectroscopy (vis-VIS-NIR) and image processing. The preprocessed vis-VIS-NIR spectral data are fused with image features after sequential projection algorithm (SPA) feature selection. Subsequently, traditional machine learning and deep learning classification models are compared, with the support vector machine (SVM) and convolutional neural network (CNN) models yielding the highest prediction rates among traditional machine learning models and deep learning models with 97.14% and 95.15% in the prediction set, respectively. The results indicate that VIS-NIR combined with image processing possesses the capability for rapid non-destructive online determination of the fermentation degree of Oolong tea. Additionally, the predictive rate of traditional machine learning models exceeds that of deep learning models in this study. This study provides a theoretical basis for the fermentation of Oolong tea.

Feng X., Wang H., Zhu Y., Ma J., Ke Y., Wang K., Liu Z., Ni L., Lin C., Zhang Y., Liu Y.

Liu X., An H., Cai W., Shao X.

Deng Z., Wang T., Zheng Y., Zhang W., Yun Y.

The development of fast, efficient, accurate, and reliable techniques and methods for food authenticity identification is crucial for food quality assurance. Traditional machine learning algorithms often have limitations when handling complex sample data, exhibiting a suboptimal performance, particularly when addressing intricate problems and in large-scale data applications. In recent years, the emergence of deep learning algorithms has heralded revolutionary breakthroughs in the field of food authenticity identification, and the ongoing deep learning developments will continue to propel advancements in this field. This review presents an overview of the deep learning algorithms and various categories of deep neural network models and structures, including the multilayer perceptron (MLP), convolutional neural network (CNN), recurrent neural network (RNN), autoencoder (AE), generative adversarial network (GAN), and attention mechanism (AM). It also summarizes the applications of these models, as well as the use of integrated models together with various analytical techniques in food authenticity. In addition, the latest developments and trends in deep learning in this field are discussed. The formidable capabilities of deep learning algorithms, in synergy with a broad array of analytical techniques, enhance the precision and efficiency of the analysis of the diverse food components. Concurrently, they have distinct advantages over traditional machine learning algorithms, showing significant potential for food authenticity identification. Although the use of deep learning still faces some challenges, with continuous technological advancements, more deep learning applications are expected to emerge in the food industry in the future to safeguard food authenticity.

Wei Y., Wen Y., Huang X., Ma P., Wang L., Pan Y., Lv Y., Wang H., Zhang L., Wang K., Yang X., Wei X.

Tea is a globally significant agricultural product, renowned for its economic and cultural value. The process of tea cultivation and production involves tea plantation management, disease control, harvesting, processing, sorting and safety and quality assessment. The quality of tea can be affected by many factors, involving variety, environment, picking and processing. Nevertheless, quality assessment of tea often relies on manual experience and specialized knowledge, which is accompanied by subjectivity and inconsistency. Furthermore, the tea production process also faces several challenges, such as pest and disease prediction and detection, supply chain monitoring and traceability. This review introduces intelligent technologies applied in tea industry, including computer vision, machine learning, spectroscopic techniques, artificial sensors, big data, internet of things, and blockchain. We summarize the progress of the application of intelligent technologies in tea industry, analyze the existing challenges and gaps, and suggest future research trends. The review is expected to provide novel insights into the application of intelligent technologies in tea industry to build a transparent, traceable, and sustainable tea industry chain. Intelligent technologies have a broad application prospect in tea industry to improve product quality, efficiency, transparency, and traceability. Particularly, combination of intelligent technologies may result in better performance. Open datasets are necessary for storage of huge amount of information. Standardization of intelligent technologies establishes a solid foundation for development of sustainable tea industry. Furthermore, transition to portable devices is the most responsive direction to tea market demands.

Wan M., Yan T., Xu G., Liu A., Zhou Y., Wang H., Jin X.

Soil available nutrients are crucial for promoting crop growth, and controlling their content is essential for increasing yield, promoting smart agriculture, and protecting the environment. Near-infrared spectroscopy technology enables the efficient and nondestructive detection of soil nutrient content in real time. However, current near-infrared spectroscopy datasets suffer from data isolation due to the inability to share data feature advantages, necessitating expensive data acquisition. To address this issue, we propose an unsupervised method for near-infrared spectral enhancement to analyse soil samples from the red soil of southern Anhui. Our proposed framework, named MAE-NIR, is a near-infrared spectral masked autoencoder that learns highly robust and generic spectral features from abundantly available public near-infrared spectral datasets. We collected near-infrared spectroscopy data from a depth of 900–1700 nm and utilized the publicly available spectral dataset LUCAS 2009 to reconstruct the spectral waveform details. This method facilitates the capture of both local and global aspects, thereby aiding subsequent downstream tasks. Several renowned regressors, such as partial least squares regression, random forest, and neural networks, are also employed to assess the effectiveness of near-infrared spectral enhancement using MAE-NIR. The spectral enhancement method based on the masked autoencoder significantly outperforms all other spectral preprocessing methods. The coefficient of determination (R2) values of the best models of available nitrogen, phosphorus, and potassium in the soil increased to 0.941, 0.926, and 0.903, respectively, which are on average 22.42%, 11.14%, and 10.35% higher than those obtained from the previous best preprocessing methods. This indicates the efficacy of using MAE-NIR to predict soil nutrients, as it effectively enhances the accuracy of nutrient content measurements.

Xue J., Liu P., Feng L., Zheng L., Gui A., Wang X., Wang S., Ye F., Teng J., Gao S., Zheng P.

Fresh leaves of Echa 1 were fixed by roller, steam/hot air and light-wave, and the effects of the three fixation methods on the chemical characteristics of straight-shaped green teas (GTs) were studied by widely targeted metabolomic analysis. 1001 non-volatile substances was identified, from which 97 differential metabolites were selected by the criteria of variable importance in projection (VIP) > 1, p < 0.05, and |log2(fold change)| > 1. Correlation analysis indicated that 14 taste-active metabolites were the major contributors to the taste differences between differently processed GTs. High-temperature fixation induces protein oxidation or degradation, γ-glutamyl peptide transpeptidation, degradation of flavonoid glycosides and epimerization of cis-catechins, resulting in the accumulation of amino acids, peptides, flavonoids and trans-catechins, which have flavor characteristics such as umami, sweetness, kokumi, bitterness and astringency, thereby affecting the overall taste of GTs. These findings provided a scientific basis for the directional processing technology of high-quality green tea.