Multi-Scale Mixed Attention Tea Shoot Instance Segmentation Model

Tea leaf picking is a crucial stage in tea production that directly influences the quality and value of the tea.Traditional tea-picking machines may compromise the quality of the tea leaves.High-quality teas are often handpicked and need more delicate operations in intelligent picking machines.Compared with traditional image processing techniques,deep learning models have stronger feature extraction capabilities,and better generalization and are more suitable for practical tea shoot harvesting.However,current research mostly focuses on shoot detection and cannot directly accomplish end-to-end shoot segmentation tasks.We propose a tea shoot instance segmentation model based on multi-scale mixed attention(Mask2FusionNet)using a dataset from the tea garden in Hangzhou.We further analyzed the characteristics of the tea shoot dataset,where the proportion of small to medium-sized targets is 89.9%.Our algorithm is compared with several mainstream object segmentation algorithms,and the results demonstrate that our model achieves an accuracy of 82%in recognizing the tea shoots,showing a better performance compared to other models.Through ablation experiments,we found that ResNet50,PointRend strategy,and the Feature Pyramid Network(FPN)architecture can improve performance by 1.6%,1.4%,and 2.4%,respectively.These experiments demonstrated that our proposed multi-scale and point selection strategy optimizes the feature extraction capability for overlapping small targets.The results indicate that the proposed Mask2FusionNet model can perform the shoot segmentation in unstructured environments,realizing the individual distinction of tea shoots,and complete extraction of the shoot edge contours with a segmentation accuracy of 82.0%.The research results can provide algorithmic support for the segmentation and intelligent harvesting of premium tea shoots at different scales.

METABOLIC AND TRANSCRIPTOME ANALYSIS REVEALS METABOLITE VARIATION AND FLAVONOID REGULATORY NETWORKS IN FRESH SHOOTS OF TEA(CAMELLIA SINENSIS)OVER THREE SEASONS

Metabolites,especially secondary metabolites,are very important in the adaption of tea plants and the quality of tea products.Here,we focus on the seasonal variation in metabolites of fresh tea shoots and their regulatory mechanism at the transcriptional level.The metabolic profiles of fresh tea shoots of 10 tea accessions collected in spring,summer,and autumn were analyzed using ultra-performance liquid chromatography coupled with quadrupole-obitrap mass spectrometry.We focused on the metabolites and key genes in the phenylpropanoid/flavonoid pathway integrated with transcriptome analysis.Multivariate statistical analysis indicates that metabolites were distinctly different with seasonal alternation.Flavonoids,amino acids,organic acids and alkaloids were the predominant metabolites.Levels of most key genes and downstream compounds in the flavonoid pathway were lowest in spring but the catechin quality index was highest in spring.The regulatory pathway was explored by constructing a metabolite correlation network and a weighted gene co-expression network.

Behavioral responses for evaluating the attractiveness of specific tea shoot volatiles to the tea green leafhopper, Empoaca vitis

The tea green leafhopper, Empoasca vitis Gothe, is one of the most serious insect pests of tea plantations in China's Mainland. Over the past decades, this pest has been controlled mainly by spraying pesticides. Insecticide applications not only have become less effective in controlling damage, but even more seriously, have caused high levels of toxic residues in teas, which ultimately threatens human health. Therefore, we should seek a safer biological control approach. In the present study, key components of tea shoot volatiles were identified and behaviorally tested as potential leafhopper attractants. The following 13 volatile compounds were identified from aeration samples of tea shoots using gas chromatography-mass spectrometry （GC-MS）： （E）-2-hexenal, （Z）-3-hexen-1- ol, （Z）-3-hexenyl acetate, 2-ethyl-1-hexanol, （E）-ocimene, linalool, nonanol, （Z）-butanoic acid, 3-hexenyl ester, decanal, tetradecane, β-caryophyllene, geraniol and hexadecane. In Y-tube olfactometer tests, the following individual compounds were identified： （E）-2- hexenal, （E）-ocimene, （Z）-3-hexenyl acetate and linalool, as well as two synthetic mixtures （called blend 1 and blend 2） elicited significant taxis, with blend 2 being the most attractive. Blend 1 included linalool, （Z）-3-hexen-l-ol and （E）-2-hexenal at a 1： 1：1 ratio, whereas blend 2 was a mixture of eight compounds at the same loading ratio： （E）-2-hexenal, （Z）- 3-hexen-l-ol, （Z）-3-hexenyl acetate, 2-penten-l-ol, （E）-2-pentenal, pentanol, hexanol and 1-penten-3-ol. In tea fields, the bud-green sticky board traps baited with blend 2, （E）-2- hexenal or hexane captured adults and nymphs of the leafhoppers, with blend 2 being the most attractive, foUowed by （E）-2-hexenal and hexane. Placing sticky traps baited with blend 2 or （E）-2-hexenal in the tea fields significantly reduced leathopper populations. Our results indicate that the bud-green sticky traps baited with tea shoot volatiles can provide a new tool for monitoring and managing the tea leafhopper.

High-efficiency tea shoot detection method via a compressed deep learning model

Achieving high-efficiency and accurate detection of tea shoots in fields are essential for tea robotic plucking. A real-time tea shoot detection method using the channel and layer pruned YOLOv3-SPP deep learning algorithm was proposed in this study. First, tea shoot images were collected and data augmentation was performed to increase sample diversity, and then a spatial pyramid pooling module was added to the YOLOv3 model to detect tea shoots. To simplify the tea shoot detection model and improve the detection speed, the channel pruning algorithm and layer pruning algorithm were used to compress the model. Finally, the model was fine-tuned to restore its accuracy, and achieve the fast and accurate detection of tea shoots. The test results demonstrated that the number of parameters, model size, and inference time of the tea shoot detection model after compression reduced by 96.82%, 96.81%, and 59.62%, respectively, whereas the mean average precision of the model was only 0.40% lower than that of the original model. In the field test, the compressed model was deployed on a Jetson Xavier NX to conduct the detection of tea shoots. The experimental results demonstrated that the detection speed of the compressed model was 15.9 fps, which was 3.18 times that of the original model. All the results indicate that the proposed method could be deployed on tea harvesting robots with low computing power to achieve high efficiency and accurate detection.