Nitrogen‑phosphorus responses and Vis/NIR prediction in fresh tea leaves

Area of Science:

• Agricultural Science
• Plant Nutrition
• Spectroscopy

Background:

• Nitrogen (N) and phosphorus (P) are crucial for tea plant growth.
• Nutrient content varies significantly across different tea plant parts.
• Limited research exists on N and P content in fresh tea leaves.

Purpose of the Study:

• To analyze nitrogen content (NC) and phosphorus content (PC) in fresh tea leaves under varying nitrogen stress.
• To develop quantitative regression models for predicting NC and PC using Vis/NIR spectroscopy and intelligent algorithms.
• To investigate the impact of preprocessing algorithms on model predictive performance.

Main Methods:

• Vis/NIR spectroscopy was employed for spectral data acquisition.
• Various intelligent algorithms were utilized for quantitative regression modeling.
• Preprocessing algorithms were evaluated for their impact on model accuracy.
• Specific models (VMDSG-D1-VCPA-IRIV-SVR for NC, VMDSG-CARS-Stacking for PC) were developed and tested.

Main Results:

• Significant differences in NC and PC were observed across different tea plant parts.
• The choice of preprocessing algorithms critically influenced model predictive performance.
• The developed VMDSG-D1-VCPA-IRIV-SVR and VMDSG-CARS-Stacking models demonstrated high predictive accuracy for NC and PC, respectively.
• Test set correlation coefficients exceeded 0.85, with RPD values greater than 1.8.

Conclusions:

• The study successfully established accurate predictive models for NC and PC in fresh tea leaves.
• These findings provide a foundation for precise fertilization strategies in tea cultivation.
• The developed spectroscopic methods enable efficient in-situ detection of nutrient status in tea plants.