The Near-infrared Spectroscopy Analysis Of Soil Organic Matter And Related Software Development

In precision agriculture, an very important part is variable rate fertilizer. To carry out variable rate fertilizer, we must get the farmland soil nutrient contents. Soil nutrients includes organic matter, nitrogen, phosphorus and potassium which are the basis of crop growth and the measurement index of soil productivity and soil fertility.Although the traditional chemical analysis method of soil nutrients has the high accuracy, it is complicated and time-consuming. By contrast, the near-infrared spectroscopy analysis technique is rapid, real-time highly efficient, lossless, non-polluting. So in the aspect of soil nutritions analysis, the near-infrared spectroscopy has a huge potential and it has becomes the hot field at home and abroad.In this paper,147soil samples are collected from Long Kang farm in Huaiyuan Country, Anhui Province. Meanwhile, the prediction model is built for soil organic matter by using the principal component regression. In the experiment, Kennard-Stone algorithm is applied to divide soil samples into calibration set and validation set. The correlation coefficients(Rp2)and root mean square error of prediction (RMSEP)are used as the indicators to choose optimal spectral dada pretreatment. The result shows that combine Savitzky-Golay and standard normal variable transformation (SNV) can get the best effect. Furthermore, this paper put forward a new method of optimizing wavelength and it turns out the prediction precision of model can be significantly improved by this method. Based on above, finally the Rp2and RMSEP of the prediction model for soil organic matter reach0.9322and0.0411%.At the same time, another innovative point of this paper is as follows.Based on the hardware platform in the laboratory for collecting the soil visible-Near-infrared spectroscopy, corresponding software is developed by using the development tool of Microsoft Visual studio2010. The software system includes the function of collecting, displaying and saving the soil spectroscopy, the absorbance and reflectivity calculation, spectral data pretreatment(such as Savitzky-Golay, SNV, first derivative and second derivative) and the prediction of soil organic matter content.