UAV hyperspectral inversion modeling of rice nitrogen content based on WOA-ELM

Abstract

Abstract: Nitrogen content is an important indicator of the growth status of the rice in cold region.  It can be obtained in time using UAV hyperspectral remote sensing technology at a regional scale.  This study is based on the remote sensing test data of the rice in precision agriculture aviation team experiment station of Shenyang Agricultural University from 2018.  The method of Sequent Projection Approach (SPA) was used to extract the effective band including 465 nm, 501 nm, 578 nm, 702 nm and 783 nm.  The extracted characteristic band is used as the input, and the inversion models of nitrogen content in rice canopy were established respectively by using the Extreme Learning Machine (ELM), and the Extreme Learning Machine by genetic algorithm (GA-ELM) and by Whale optimization algorithm (WOA-ELM).  The results showed that the precision of the nitrogen content of rice based on the WOA-ELM was better than the inversion model established by ELM and GA-ELM.  The R² of training data is 0.887, the R² of test data is 0.880, the RMSE of training data is 0.269, the RMSE of test data is 0.284.  This study provides a certain data support and application basis for the diagnosis of rice nitrogen content in cold region by UAV hyperspectral remote sensing technology in northeastChina.

Keywords: UAV, nitrogen content, rice in cold region, hyperspectral, WOA-ELM

DOI: 10.33440/j.ijpaa.20190202.39.

 

Citation: Yu F H, Du W, Guo Z H, Zhou C X, Wang D K, Xu T Y.  UAV hyperspectral inversion modeling of rice nitrogen content based on WOA-ELM.  Int J Precis Agric Aviat, 2019; 2(2): 43–48.