The Research On Processing Method Of NIR Diffuse Reflectance Spectral Signal Based On Wavelet Transform

In people's daily production and life, rapid and non-destructive measuring of grain composition has important significance. Near infrared spectral analysis technology is a high efficient, fast, low cost, no damage and green environmental protection's new technology, it can be used for site fast detection and real-time on-line analysis.In the thesis, firstly, the development and application of the near-infrared spectroscopy analysis and Wavelet transform technology are introduced at home and abroad, the basic principles of the near-infrared spectroscopy analysis and Wavelet transforms are studied; secondly, This article systematically introduce a lot of traditional Pretreatment methods and several calibration mode of Near-infrared spectra; lastly, this article has developed its own set of wheat near-infrared spectroscopy instruments, and in-depth analysis and research in the wavelet threshold denoising and the wavelet coefficients of the iPLS mode method, we analysis wavelet threshold denoising effect by the Signal Noise Ratio (SNR),Root Mean Square Error (RMSE) and modeling results respectively, and the modeling result of Wavelet coefficients (Cd4,Cd5,Cd7) iPLS modeling is more precise and more predictable.