The Judgement on Lack of Nitrogen in Rice Based on SVM Algorithm

The intelligent judgment of nitrogen content in rice has great significance to its healthy growth. Here use soil less culture technique to cultivate rice with different nitrogen levels. In various stages, field spec is used to acquire rice canopy spectra, meanwhile using AA3 continuous flow analyzer to measure nitrogen level in specimen of rice leaves, which constitutes a sample database about rice with different nitrogen content. Samples in the sample library is tested and modeled by use of Least Squares Support Vector Machines (LS-SVM) method. At last, predict the classification of the test samples. This test shows that, the eventual recognition accuracy whether are deficient in nitrogen can reach 95%. It suggests that, support vector machines (SVM) can be used for the judgment whether the rice is deficient in nitrogen or not.