Application of the chlorophyll fluorescence ratio in evaluation of paddy rice nitrogen status

Abstract

In this research, laser-induced fluorescence (LIF) technique combined with back-propagation neural network (BPNN) was employed to analyse different nitrogen (N) fertilization levels in paddy rice. Leaf fluorescence characteristics (FLCs) were measured by using the LIF system built in our laboratory and exhibited different FLCs with different nitrogen fertilization levels. The correlation between fluorescence intensity ratios (F685/F460, F735/F460 and F735/F685) and the dose of N fertilization was established and analysed. Then, the BPNN algorithm was utilized to validate that the different N fertilization levels can be classified based on the three FLCs. The overall identification accuracies of 2014 and 2015 were 90% and 92.5%, respectively. Experimental results demonstrated that the three FLCs with the help of multivariate analysis can be served as a helpful tool in the evaluation of paddy rice N fertilization levels. Besides, this study can also provide guidance for the selection of LIF Lidar channels in the following research.

Keywords

• fluorescence characteristics
• remote sensing
• nutrient stress
• oryza sativa
• machine learning