Caffeine is an important component that determines the quality of tea, and its rapid estimation is very much needed for the industry. In this pursuit, a near‐infrared (NIR) spectroscopy‐based technique for the estimation of caffeine is developed and presented in this paper. On the basis of responses of the different bonds present in caffeine, four specific wavelength windows—(a) 1075 to 1239.5 nm (C―H stretch second overtone); (b) 1339.25 to 1440.75 nm (C―H stretch and C―H deformation); (c) 1640.25 to 1700 nm (C―H stretch first overtone, ═CH & amp; ―CH3 asymmetric); and (d) 900 to 1700 nm (whole range of the spectrometer)—were analyzed in details for model development and to obtain the effective wavelength (EW). Five different preprocessing techniques followed by two regression techniques—(a) the partial least‐squares (PLS) and (b) the support vector regression (SVR) were implemented on raw data for analysis. Comparing all the models, the wavelength band of 1075 to 1239.5 nm and 1339.25 to 1440.75 nm were found to produce satisfactory results. The best discrimination result was obtained using the combination of standard normal variate (SNV) preprocessing with SVR at the 1075 to 1239.5 nm wavelength region. The SVR regression with 105 samples in the training set and 15 samples in the testing set resulted in the performance parameters as RMSECV = 0.134, RMSEP = 0.069, rcv2 = 0.869, rp2 = 0.65, and RPD = 5.626 at 1075 to 1239.5 nm, whereas the PLS model produced the best RMSECV = 0.287, RMSEP = 0.077, rcv2 = 0.637, rp2 = 0.675, and RPD = 5.218 at 1339.25 to 1440.75‐nm wavelength band.

This paper reports on the development of an integrated leaf quality inspecting system using near infrared reflectance (NIR) spectroscopy for quick and in situ estimation of total polyphenol (TP) content of fresh tea leaves, which is the most important quality indicator of tea. The integrated system consists of a heating system to dry the fresh tea leaves to the level of 3-4% moisture, a grinding and sieving system fitted with a 250 micron mesh sieve to make fine powder from the dried leaf. Samples thus prepared are transferred to the NIR beam and TP is measured instantaneously. The wavelength region, the number of partial least squares (PLS) component and the choice of preprocessing methods are optimized simultaneously by leave-one-sample out cross-validation during the model calibration. In order to measure polyphenol percentage in situ, the regression model is developed using PLS regression algorithm on NIR spectra of fifty-five samples. The efficacy of the model developed is evaluated by the root mean square error of cross-validation, root mean square error of prediction and correlation coefficient (R 2 ) which are obtained as 0.1722, 0.5162 and 0.95, respectively.Keywords Polyphenol Á Fine leaf count Á Near infrared reflectance (NIR) spectroscopy Á Folin Ciocalteu method Á Partial least squares (PLS) Á Preprocessing Á Root mean square error of cross-validation (RMSECV)

No abstract

No abstract

The quality of tea liquor primarily depends on the quality of inbound fresh leaves to a tea factory. We report here the development of an industrial setup using diffuse reflectance near-infrared spectroscopy for quick and on-spot estimation of total polyphenol content, which is one of the major quality indicators of tea. In the inspection system, freshly plucked tea leaves are dried with a heating system designed in such a way that the moisture content is reduced to around 3-4%. The dried leaves are then ground and sieved through a 250 mm mesh. Samples thus produced are placed automatically before the near-infrared beam and assessment of total polyphenol is done instantaneously.