Reserch Of Living Tea Based On Near Infrared Spectral

With the development of science and technology, among the many substances contained in tea, of the natural active ingredients, Anti-oxidation, anti-cancer and hypoglycemic, enhance immunity and anti-aging effects were exhumed, which expand the size of tea market. China is the world’s largest tea producer and tea occupies a very important position in China’s economic crops. At present, the tea garden management, in order to grasp the growth status of tea, by the appearance of the diagnostic manual experience, sensory evaluation of fresh tea acceptance grading and pricing is a common method, but this method greatly influenced by subjective factors and lacks reliable scientific basis, and takes up a lot of manpower and material resources, Moreover the evaluation results is susceptible to subjective factors, there is a deviation. Therefore, the establishment of a real-time, fast, high-precision detection of fresh tea quality is imperative.On the basis of the comprehensive analysis of content in fresh tea material and high-band spectral characteristics relevance, this article was first proposed to extract the spectral data in the fresh tea -related material and high spectral content online testing were detected in tea by high performance liquid chromatography.In this study, the detection of targets is the content of polyphenols, amino acids. And combined with the quality of the research stoichiometry and quantitative analysis of near infrared spectroscopy to start testing the quality of tea, choosing from pre-treatment method mainly to optimize the analysis and discussion, and related modeling methods, provides a new basis for fresh tea tea quality online non-destructive testing. In this paper, the innovative design is the near-infrared line detection of fresh content in tea and tea polyphenols and amino acids, and without destroy to organized tea growing collection of spectral data with a more precise control of the guidance, and takes the quality of tea leaves and a bud and two buds of tea as an important indicator of material content, establish relevant linear regression equation to provide a basis for detecting a wide range of tea.In this study, a total of 108 tea samples of synchronization using near infrared spectroscopy its real-time field data collection were collected, at the same time, the tea picking area, after preservation treatment, were detected by HPLC sample tea polyphenols content and amino acids. Comparing the first derivative, second derivative, smoothing and other pretreatment effect, and the effect of window size for each method of model accuracy, combined pretreatment identify polyphenols and amino acids to obtain the best pretreatment methods. Experiments show that the size of the window 108 samples polyphenols and amino acid composition were optimal prediction models 95 and 105 of its smoothed when the first derivative and smoothing treatment combinations, which were the lowest variance model predicts, respectively,2.553 and 3.013.In this paper, after pretreatment optimized, using principal component analysis of spectral data selected characteristic bands combine high performance liquid chromatography measured the polyphenol content of amino acids and built a calibration models. Its principal component analysis feature bands of tea polyphenols and amino acids, respectively 821nm,765nm,940nm,520nm and 500-512nm,702-709um,754nm,726-732nm. Through a linear, exponential, power function, polynomial regression, least squares method and multiple linear regression calibration model established, and analyzed predictive accuracy. The results and the accuracy of the calibration model prediction model established by the least squares method highest polyphenols and polyphenol content prediction correlation of 0.97 and 0.96, respectively, to predict the mean square error of 1.96 and 1.11 respectively. The forecasts higher rate was able to achieve the purpose of forecasting.