Fast Analyzer Principle And Design Of Near-infrared Technology-based Tea Ingredients

The near-infrared spectral analysis technology has some characteristic, such as fast, stable, accurate, abstemious in energy sources, low cost. etc. It substitutes for the normal physicochemical analysis ways quickly, and it has been widely used in component analysis in food and farm produce. It was considered as one of the most important technologies in analytical chemistry and analytical instrumentation in PITTCON.Chemometrics and near-infrared spectral (NIR) analysis technology were applied, this paper was mainly carried about the studied of the theory and design of the fast analytical instrumentation of components tea.The near-infrared absorption spectrum characteristic was analyzed in detail, the relation between the spectrum and the components was discussed, and the non-linearity of them was proved by the correlation coefficient and partial least squares method (PLS), the models were improved after the radial basis function (RBF) neural network adopted. In order to improve the ratio of ability and cost of the system, this paper proposed the method of characteristic wavelengths chosen based on genetic algorithm (GA), and the characteristic wavelengths of tea were chosen. The results shown that the GA model was better than the PLS and BRF neural network model. This paper finished some design of the circuits, after the compare of the structure about a few kinds of NIR spectrum instrument, the design of the NIR spectrum instrument used the filter and characteristic wavelengths of tea was proposed.