Visualization of Moisture Distribution in Stacked Tea Leaves on Process Flow Line Using Hyperspectral Imaging

Area of Science:

• Agricultural Engineering
• Food Science
• Spectroscopy

Background:

• Tea quality is significantly impacted by moisture content distribution during processing.
• Optimizing tea processing requires effective visualization of moisture levels in stacked tea leaves.

Purpose of the Study:

• To evaluate moisture content and distribution in stacked tea leaves using hyperspectral imaging (HSI) and machine learning.
• To develop an accurate model for predicting moisture content in green tea products during processing.

Main Methods:

• Utilized hyperspectral imaging (HSI) technology combined with machine learning algorithms.
• Developed a spectral quantitative determination model for moisture content assessment.
• Applied the model to West Lake Longjing and Tencha green tea during processing.

Main Results:

• Achieved high accuracy (Rp2 > 0.940) in the developed spectral quantitative determination model.
• Demonstrated strong generalization ability of the model across different tea types.
• Visualized non-uniform moisture distribution, with higher levels at leaf tips and petioles compared to the leaf surface.

Conclusions:

• HSI and machine learning offer a novel solution for real-time moisture monitoring in stacked tea leaves.
• The findings enable optimization of processing parameters to ensure consistent tea product quality.
• Future work should focus on model transferability and application to diverse tea varieties.