Quality Assessment of Jasmine Tea Combining Computer Vision and Color Difference Analysis.

Area of Science:

• Food Science
• Agricultural Engineering
• Analytical Chemistry

Background:

• Jasmine tea quality assessment is challenging due to the ephemeral nature of jasmine blossoms.
• Traditional methods struggle to consistently evaluate the complex sensory attributes of jasmine tea.

Purpose of the Study:

• To develop an objective quality assessment method for jasmine tea using computer vision and color difference technology.
• To establish correlations between visual parameters and key chemical components (free amino acids, polyphenols, caffeine).

Main Methods:

• Utilized computer vision and color difference technology for objective tea analysis.
• Correlated color parameters (a*, ∆E, I) with sensory evaluations and chemical constituents.
• Developed and compared standard backpropagation (BP) and genetic algorithm-optimized backpropagation (GA-BP) neural networks for predictive modeling.

Main Results:

• Significant correlations were found between sensory evaluation and free amino acids, polyphenols, and caffeine.
• Tea polyphenols strongly correlated with a* (r=0.69) and ∆E (r=-0.78).
• Caffeine content correlated with I (r=0.64), a* (r=0.50), and ∆E (r=-0.57).
• The GA-BP model achieved higher accuracy in predicting polyphenol and caffeine content compared to the BP model.

Conclusions:

• Computer vision and color difference technology provide an efficient and novel approach for jasmine tea quality assessment.
• The GA-BP neural network model offers superior predictive accuracy for key chemical components.
• This method addresses the limitations of traditional quality assessment methods for jasmine tea.