Analysis And Modeling Of NIR Spectrum Based On Intelligent Algorithm

Race is one of the most important grain plants. Quickly and non-damage to measure the component of wheat is very important in the field of race's purchase and production. The near-infrared spectroscopy analysis technique is efficient, rapid, low cost, noninvasive and not destroying environment. It is not only suitable for laboratory analysis, but also for in-field fast and real-time on-line analysis. But foreign Near-infrared spectroscopy analysis software can't analyze our country's sample efficiently. So, in the dissertation, an in-depth study is carried out by synthesizing multi-disciplinary knowledge on the pre-treatment and modeling of race's Near-infrared spectroscopy.This article systematically analyzes several traditional pretreatment methods for Near-infrared spectra, and discusses the character and the field they can be used in. For the advantage of wavelet transform in denoising and data compression, we choose wavelet transform to denoise and compress the data of Near-infrared spectra. Based on the shortcoming of wavelet transform that it can't analyze the relations between spectra and contents, a method based on rough set is presented. Several reduction algorithms are studied in this article. Based on the practical situation of the program, the algorithm base on discernibility matrix is chosen as the algorithm used in computation through programming.Based on the linear features of traditional spectroscopy analysis technique, ANN and GA-Fuzzy method are presented in the dissertation. Because of the non-linear approximation of BP network the author chosen BP network as the modeling method of the spectra. But for it can't interpret the model effectively, a new method (GA-Fuzzy model) is presented. To get the optimal model, the method optimizes the parameters of T-S fuzzy model use Coevolutionary genetic algorithm. Finally it is proved that GA-Fuzzy model is better than BP network model through the experiment.The study not only offers several suitable methods for near-infrared spectroscopy analysis and quantitative analysis of content of race but also offers a new direction to the theoretic development and practical application of near-infrared spectroscopy analysis technique.