Medicinal plants have incredibly high economic value, and a practical evaluation of their quality is the key to promoting industry development. The deep learning model based on residual convolutional neural network (ResNet) has the advantage of automatic extraction and the recognition of Fourier transform near-infrared spectroscopy (FT-NIR) features. Models are difficult to understand and interpret because of unknown working mechanisms and decision-making processes. Therefore, in this study, artificial feature extraction methods combine traditional partial least squares discriminant analysis (PLS-DA) and support vector machine (SVM) models to understand and compare deep learning models. The results show that the ResNet model has significant advantages over traditional models in feature extraction and recognition. Secondly, preprocessing has a great impact on the feature extraction and feature extraction, and is beneficial for improving model performance. Competitive adaptive reweighted sampling (CARS) and variable importance in projection (VIP) methods screen out more feature variables after preprocessing, but the number of potential variables (LVs) and successive projections algorithm (SPA) methods obtained is fewer. The SPA method only extracts two variables after preprocessing, causing vital information to be lost. The VIP feature of traditional modelling yields the best results among the four methods. After spectral preprocessing, the recognition rates of the PLS-DA and SVM models are up to 90.16% and 88.52%. For the ResNet model, preprocessing is beneficial for extracting and identifying spectral image features. The ResNet model based on synchronous two-dimensional correlation spectra has a recognition accuracy of 100%. This research is beneficial to the application development of the ResNet model in foods, spices, and medicinal plants.

Due to a similar plant morphology in the majority of Zingiberaceae spices, substitution and adulteration frequently take place during the sales process. Therefore, it is important to analyze the metabolites and species classification of different Zingiberaceae spices. This study preliminarily explored the differences in the metabolites in thirteen Zingiberaceae spices through untargeted gas chromatography–mass spectrometry (GC–MS) and combined spectroscopy, establishing models for classifying different Zingiberaceae spices. On one hand, a total of 81 metabolites were successfully identified by GC–MS. Thirty-seven differential metabolites were screened using variable important in projection (VIP ≥ 1). However, the orthogonal partial least squares discriminant analysis (OPLS-DA) model established using GC–MS data only explained about 30% of the variation. On the other hand, the partial least squares discriminant analysis (PLS-DA) models with three spectral data fusion strategies were compared, and their classification accuracy reached 100%. Among them, the mid-level data fusion model based on latent variables had the best performance. This study provides a powerful tool for distinguishing different Zingiberaceae spices and assists in reducing the occurrence of substitution and adulteration phenomena.

The study used headspace solid-phase microextraction coupled with gas chromatography-time-of-flight mass spectrometry (HS–SPME–GC–TOF–MS) to analyze volatile compounds in leaves and fruits of Amomum tsaoko, Amomum paratsaoko, and Amomum koenigii. The composition and aroma of distinct metabolites were analyzed using multivariate statistical methods. A total of 564 volatile compounds were identified from three species of the genus Amomum, which were further divided into nine categories: terpenoids, carboxylic acids, alcohols, hydrocarbons, aldehydes, ketones, phenols, ethers, and other compounds. Terpenoids and alcohols were the most abundant. The content and types of compounds vary in A. tsaoko, A. paratsaoko, and A. koenigii, so mixing or substituting them is not advisable. We selected 45 metabolites based on the criteria of the variable importance in projection values (VIP > 1.5) and one-way ANOVA (p < 0.05). The top 19 metabolites with the most significant VIP values were chosen. Interestingly, (Z)-2-decenal was only found in Amomum koenigii, while nitroethane and nonanal were only present in cultivated A. tsaoko. Additionally, linalool, cineole, and (D)-limonene were the main components affecting the aroma of three species of the genus Amomum. The volatile components identified in this study provide a theoretical basis for analyzing the unique flavor of A. tsaoko, A. paratsaoko, and A. koenigii.

The fruit of Lanxangia tsao - ko (LT), a widely used spice, is highly perishable during storage and transportation making it essential to develop effective preservation methods. This study employed multi-omics techniques to evaluate the effects of three different treatments on the quality of LT during short-term storage. Results demonstrated that, compared to the tap water control group (TW), the hot water treatment group (HW), carboxymethyl chitosan treatment group (CMC), and hydrogen peroxide‑silver ion solution treatment group (SP) significantly suppressed the increase in color change intensity (ΔE), browning index (BI), weight loss rate (W%), and rot rate (R%) of LT. Furthermore, these treatments promoted the accumulation of beneficial flavor substances, while reducing the level of unfavorable ones. Both CMC and SP were effective in inhibiting the increase of LT's respiration intensity and the relative abundance of pathogenic microorganisms like Rhizopus microsporus and Rhizopus arrhizus. However, HW led to an abnormal decrease in respiration intensity and increases the relative abundance of certain pathogenic microorganisms. In summary, although the temperature setting in the hot water treatment group was not ideal, hot water treatment still significantly inhibited the spoilage and discoloration phenomenon. Both CMC and SP can effectively extend the storage and transportation period of LT in all aspects. This study provides a new idea for the preservation of LT.