An orchard variable-rate sprayer applies the appropriate amount of plant protection products only where they are needed based on detection data from advanced sensors, a system that has attracted increasing attention. The latest developments in the detection unit, variable control unit, and signal-processing algorithm of the variable-rate sprayer are discussed. The detection of target position and volume is realized with an ultrasonic sensor, a laser scanning sensor, or other methods. The technology of real-time acquisition of foliage density, plant diseases and pests and their severity, as well as meteorological parameters needs further improvements. Among the three variable-flow-rate control units, pulse width modulation was the most widely used, followed by pressure-based, and variable concentration, which is preliminarily verified in the laboratory. The variable air supply control unit is tested both in the laboratory and in field experiments. The tree-row-volume model, the leaf-wall-area model, and the continuous application mode are widely used algorithms. Advanced research on a variable-rate sprayer is analyzed and future prospects are pointed out. A laser-based variable-rate intelligent sprayer equipped with pulse width modulation solenoid valves to tune spray outputs in real time based on target structures may have the potential to be successfully adopted by growers on a large scale in the foreseeable future. It will be a future research direction to develop an intelligent multi-sensor-fusion variable-rate sprayer based on target crop characteristics, plant diseases and pests and their severity, as well as meteorological conditions while achieving multi-variable control.

The quality and flavor of tea leaves are significantly influenced by chemical composition, with the content of free amino acids serving as a key indicator for assessing the quality of Tencha. Accurately and quickly measuring free amino acids during tea processing is crucial for monitoring and optimizing production processes. However, traditional chemical analysis methods are often time-consuming and costly, limiting their application in real-time quality control. Hyperspectral imaging (HSI) has shown significant effectiveness as a component detection tool in various agricultural applications. This study employs VNIR-HSI combined with machine learning algorithms to develop a model for visualizing the total free amino acid content in Tencha samples that have undergone different processing steps on the production line. Four pretreating methods were employed to preprocess the spectra, and partial least squares regression (PLSR) and least squares support vector machine regression (LS–SVR) models were established from the perspectives of individual processes and the entire process. Combining competitive adaptive reweighted sampling (CARS) and variable iterative space shrinkage approach (VISSA) methods for characteristic band selection, specific bands were chosen to predict the amino acid content. By comparing modeling evaluation indicators for each model, the optimal model was identified: the overall model CT+CARS+PLSR, with predictive indicators Rc² = 0.9885, Rp² = 0.9566, RMSEC = 0.0956, RMSEP = 0.1749, RPD = 4.8021, enabling the visualization of total free amino acid content in processed Tencha leaves. Here, we establish a benchmark for machine learning-based HSI, integrating this technology into the tea processing workflow to provide a real-time decision support tool for quality control, offering a novel method for the rapid and accurate prediction of free amino acids during tea processing. This achievement not only provides a scientific basis for the tea processing sector but also opens new avenues for the application of hyperspectral imaging technology in food science.

Color is a key indicator for evaluating the quality of tea during processing; various processing procedures can significantly affect the content of fat-soluble pigments of tea, which in turn affects the color and quality of finished tea. Therefore, there is an urgent demand for the fast, non-destructive detection of pigments of stacked tea during processing. This paper presents the use of hyperspectral imaging technology (HSI), combined with machine learning algorithms, to detect chlorophyll a, chlorophyll b, and carotenoids in stacked matcha tea during processing. Firstly, a quantitative relationship between HSI data of tea and their pigment contents was developed based on regression analysis, and the results showed that exceptional prediction performance was achieved by the partial least squares regression (PLSR) algorithm combined with the feature band algorithm of competitive adaptive reweighting (CARS), and the R_p² values of detection models of chlorophyll a, chlorophyll b and carotenoids were 0.90465, 0.92068 and 0.62666, respectively. Then, these quantitative detection models were extended to each pixel in hyperspectral images, achieving point-by-point prediction of pigment components, so the distribution of pigments of stacked tea leaves during processing procedures was successfully visualized on the processing line in situ. By integrating a hyperspectral imaging system into the real-world environment, operators can monitor pigment levels in real time and thus dynamically adjust processing parameters based on real-time data. This study enhances pigment detection efficiency in tea processing, supports process optimization, and aids in quality control.

As one of the six major traditional tea types in China, white tea’s quality formation is primarily influenced by the withering process. However, traditional methods for monitoring withering fail to achieve precise and stable control of moisture content. To address this issue, a total of 650 samples were collected at 13 withering time points (0–36 h), and the dataset was split into training and test sets at a 7:3 ratio. This study proposes a PRXBoost ensemble model for quantitative detection of withered white tea, which integrates data augmentation and intelligent algorithms. The ensemble model uses a Bagging-based weighted integration technique to combine Partial Least Squares Regression (PLSR), Ridge, and Extreme Gradient Boosting (XGBoost) models, and it conducts an in-depth analysis of the decision-making process within the PRXBoost model. First, the effectiveness of the data augmentation strategy and the superiority of the gradient descent algorithm are verified through pre-modeling based on the PLSR model and hyperparameter pre-search using the XGBoost model, respectively. Additionally, the Bayes algorithm is employed to optimize the weights of the sub-models, further enhancing the overall predictive performance. The results show that the PRXBoost model achieved the best performance among the compared models on the test set, with R² = 0.854 and RMSE = 0.080, exceeding the highest R² of a single model by 6%. These results indicate that PRXBoost provided improved predictive performance for moisture estimation within the current dataset. Finally, the SHapley Additive exPlanations (SHAP) algorithm is used to analyze the influence of each input feature on the prediction results, successfully identifying the 1916 nm and 1453 nm spectral bands as significant influencers of the prediction outcomes. These results suggest that the proposed model can support rapid, non-destructive monitoring of moisture evolution and provide actionable information for withering endpoint decision control.

Amid global freshwater scarcity and soil salinization, brackish irrigation is a potential alternative, yet its effects under low-leaching soilless systems remain unclear. We tested brackish irrigation (30 mmol L⁻¹ NaCl; EC ≈ 4.8 dS m⁻¹, including fertilizer) on lettuce (Lactuca sativa L.) grown in peat–perlite substrates with non-saline (CK), mildly saline (M), and moderately–severely saline (S) initial salinity. Substrate moisture and bulk electrical conductivity (EC_b) were monitored at upper, middle, and deep layers with multi-depth sensors; lettuce physiological and growth traits were measured. Under negligible drainage, salt moved downward promptly after irrigation in CK, accumulated at the surface in M, and remained high with spatiotemporal variability in S. Brackish irrigation had minimal effects on biomass and water use efficiency in CK and M, but significantly reduced both in S. These findings support tailoring brackish irrigation to initial salinity severity and motivate future work to measure drainage and calibrate EC indices to establish operational thresholds.