Pest infestations in tea gardens are one of the common issues encountered during tea cultivation. This study introduces an improved YOLOv8 network model for the detection of tea pests to facilitate the rapid and accurate identification of early-stage micro-pests, addressing challenges such as small datasets and the difficulty of extracting phenotypic features of target pests in tea pest detection. Based on the original YOLOv8 network framework, this study adopts the SIoU optimized loss function to enhance the model’s learning ability for pest samples. AKConv is introduced to replace certain network structures, enhancing feature extraction capabilities and reducing the number of model parameters. Vision Transformer with Bi-Level Routing Attention is embedded to provide the model with a more flexible computation allocation and improve its ability to capture target position information. Experimental results show that the improved YOLOv8 network achieves a detection accuracy of 98.16% for tea pest detection, which is a 2.62% improvement over the original YOLOv8 network. Compared with the YOLOv10, YOLOv9, YOLOv7, Faster RCNN, and SSD models, the improved YOLOv8 network has increased the mAP value by 3.12%, 4.34%, 5.44%, 16.54%, and 11.29%, respectively, enabling fast and accurate identification of early-stage micro pests in tea gardens. This study proposes an improved YOLOv8 network model based on deep learning for the detection of micro-pests in tea, providing a viable research method and significant reference for addressing the identification of micro-pests in tea. It offers an effective pathway for the high-quality development of Yunnan’s ecological tea industry and ensures the healthy growth of the tea industry.

To facilitate the realization of automated tea picking and enhance the speed and accuracy of tea leaf grading detection, this study proposes an improved YOLOv8 network for fresh tea leaf grading recognition. This approach integrates a Hierarchical Vision Transformer using Shifted Windows to replace segments of the original YOLOv8’s network architecture, thereby alleviating the computational load of dense image processing tasks and reducing computational expenses. The incorporation of an Efficient Multi-Scale Attention Module with Cross-Spatial Learning serves to attenuate the influence of irrelevant features in complex backgrounds, which in turn, elevates the model’s detection Precision. Additionally, the substitution of the loss function with SIoU facilitates a more rapid model convergence and a more precise pinpointing of defect locations. The empirical findings indicate that the enhanced YOLOv8 algorithm has achieved a marked improvement in metrics such as Precision, Recall, F1, and mAP, with increases of 3.39%, 0.86%, 2.20%, and 2.81% respectively, when juxtaposed with the original YOLOv8 model. Moreover, in external validations, the FPS enhancements over the original YOLOv8, YOLOv5, YOLOX, Faster RCNN, and SSD deep-learning models are 6.75 Hz, 10.84 Hz, 12.79 Hz, 28.24 Hz, and 21.57 Hz, respectively, and the mAP improvements in practical detection are 2.79%, 2.92%, 3.08%, 7.07%, and 3.84% respectively. The refined model not only ensures efficient and accurate tea-grading recognition but also boasts high recognition rates and swift detection capabilities, thereby establishing a foundation for the development of tea-picking robots and tea quality grading devices.

As a perennial plant, the nutrient supply for tea bushes is predominantly dependent on the soil. Yunnan tea plantations exhibit significant topographic slope variations, yet the combined impact of slope positions on soil chemistry and microbial communities remains unexplored. This study investigated soil chemical properties and microbial community structures across three distinct slope areas within a single tea plantation. The results showed that the contents of organic matter (OM), total nitrogen (TN), and available nutrients (AN) at the top of the slope (TS) were significantly higher than those at the foot of the slope (FS) (p < 0.001), while the cation exchange capacity (CEC) and total potassium (TK) reached peak levels in the middle of the slope (MS), with FS having the lowest nutrient levels. Redundancy analysis (RDA) indicated that bacterial communities were primarily influenced by TK, magnesium (Mg), CEC, total phosphorus (TP), and pH, whereas fungal communities were mainly regulated by TK, Mg, and CEC, highlighting the role of soil chemical properties in shaping microbial diversity and distribution. Bacterial composition showed no significant slope-related differences, but fungal communities varied notably at the family/genus levels. MS exhibited the highest microbial network complexity, suggesting stronger species interactions. Bacterial metabolic functions and fungal trophic modes were conserved across regions, indicating functional stability independent of structural changes. This study reveals slope-driven soil-microbial dynamics in Yunnan tea plantations, offering insights into microbial assembly and adaptation under topographic gradients. These findings support precision fertilization, ecological conservation, and the sustainable management of slope tea plantations.

Understanding the correlation between soil chemical properties and tea quality is essential for the comprehensive management of ancient tea gardens. However, the specific links between these factors in ancient tea gardens remain underexplored. This study analyzes the soil chemical properties of four distinct research regions in Nanhua County to explore their effects on key chemical components in ancient tea garden teas, providing a scientific basis for improving the quality of ancient tea garden teas through soil management. Employing high performance liquid chromatography (HPLC) and inductively coupled plasma mass spectrometry (ICP-MS), the chemical components of tea and the chemical properties of the soil were meticulously quantified. Following these measurements, the integrated fertility index (IFI) and the potential ecological risk index (PERI) were evaluated and correlation analysis was conducted. The results revealed that ancient tea garden tea quality is closely linked to soil chemical properties. Soil’s total nitrogen (TN), total sulfur (TS), and available potassium (AK) negatively correlate with tea’s catechin gallate (CG) component and AK also with polyphenols. Most other soil properties show positive correlations with tea components. The research also evaluated soil heavy metals’ IFI and PERI. IFI varied significantly among regions. Hg’s high pollution index indicates ecological risks; Cd in Xiaochun (XC) region poses a moderate risk. PERI suggests moderate risk for XC and Banpo (BP), with other areas classified as low risk. Implementing reasonable fertilization and soil amelioration measures to enhance soil fertility and ensure adequate supply of key nutrients will improve the quality of ancient tea gardens. At the same time, soil management measures should effectively control heavy metal pollution to ensure the quality and safety of tea products. Insights from this study are crucial for optimizing soil management in ancient tea gardens, potentially improving tea quality and sustainability.

To investigate the variation in flavonoids content in ancient tree sun–dried green tea under abiotic stress environmental conditions, this study determined the flavonoids content in ancient tree sun−dried green tea and analyzed its correlation with corresponding factors such as the age, height, altitude, and soil composition of the tree. This study uses two machine−learning models, Least Absolute Shrinkage and Selection Operator (LASSO) regression and Cox regression, to build a predictive model based on the selection of effective variables. During the process, bootstrap was used to expand the dataset for single−factor and multi−factor comparative analyses, as well as for model validation, and the goodness−of−fit was assessed using the Akaike information criterion (AIC). The results showed that pH, total potassium, nitrate nitrogen, available phosphorus, hydrolytic nitrogen, and ammonium nitrogen have a high accuracy in predicting the flavonoids content of this model and have a synergistic effect on the production of flavonoids in the ancient tree tea. In this prediction model, when the flavonoids content was >6‰, the area under the curve of the training set and validation set were 0.8121 and 0.792 and, when the flavonoids content was >9‰, the area under the curve of the training set and validation set were 0.877 and 0.889, demonstrating good consistency. Compared to modeling with all significantly correlated factors (p < 0.05), the AIC decreased by 32.534%. Simultaneously, a visualization system for predicting flavonoids content in ancient tree sun−dried green tea was developed based on a nomogram model. The model was externally validated using actual measurement data and achieved an accuracy rate of 83.33%. Therefore, this study offers a scientific theoretical foundation for explaining the forecast and interference of the quality of ancient tree sun−dried green tea under abiotic stress.

Catechins are pivotal determinants of tea quality, with soil environmental factors playing a crucial role in the synthesis and accumulation of these compounds. To investigate the impact of changes in tea garden soil environments on the catechin content in sun-dried tea, this study measured the catechin content in soil samples and corresponding tea leaves from Nanhua, Yunnan, China. By integrating the variations in catechin content with those of 17 soil factors and employing COX regression factor analysis, it was found that pH, organic matter (OM), fluoride, arsenic (As), and chromium (Cr) were significantly correlated with catechin content (p < 0.05). Further, using the LASSO regression for variable selection, a model named LCLN-CA was constructed with four variables including pH, OM, fluoride, and As. The LCLN-CA model demonstrated high fitting accuracy with AUC values of 0.674, 0.784, and 0.749 for catechin content intervals of CA ≤ 10%, 10% < CA ≤ 20%, and 20% < CA ≤ 30% in the training set, respectively. The validation set showed AUC values of 0.630, 0.756, and 0.723, respectively, indicating a well-calibrated curve. Based on the LCLN-CA model and the DynNom framework, a visual prediction system for catechin content in Yunnan sun-dried tea was developed. External validation with a test dataset achieved an $Accuracy$ of 0.870. This study explored the relationship between soil-related factors and variations in catechin content, paving a new way for the prediction of catechin content in tea and enhancing the practical application value of artificial intelligence technology in agricultural production.