Key points are not available for this paper at this time.
The geographical origin of tea significantly influences its flavor and quality. This paper proposes an intelligent method for identifying the origin of Xihu Longjing (XHLJ) tea using dynamic time-resolved colorimetric sensor array (CSA) combined with machine learning. By employing a sixteen-indicator CSA, dynamic time-resolved fingerprint representing the interactions between the volatile organic compounds (VOCs) in XHLJ tea and non-XHLJ Longjing tea was acquired. The inherent aroma chemistry variations shape the unique sensor response patterns of the array. The dynamic time-resolved fingerprint data from different tea samples were analyzed using four machine learning methods, including partial least squares discriminant analysis (PLS-DA), linear discriminant analysis (LDA), support vector machine (SVM), and convolutional neural network (CNN). CNN - based method achieved 95.60 % test set accuracy, showing high recognition and stability in distinguishing XHLJ tea from other Longjing teas. The integration of time-dependent CSA response with deep learning enables intelligent discrimination of subtle aroma differences driven by terroir. For identifying sub-regions within the XHLJ production areas, SVM achieved 95.65 % test set accuracy. This work provides a method for the geographical origin of XHLJ tea that is faster and more convenient than traditional methods, offering a cost-effective and efficient approach for tea quality assessment.
Luo et al. (Wed,) studied this question.