Parameter research of using UAV-based visible spectral analysis technology in dynamical diagnosis of nitrogen status of summer maize

Objectives Unmanned aerial vehicle (UAV) shows superiority in obtaining canopy information related to crops growth, nutrition and plant protection recently. For the practical application of this method in the diagnosis of dynamical nitrogen (N) status, a serious of visible spectrum parameters are need to set up at different stages of summer maize.

Methods Field experiment was conducted from June to October in 2015 at the Quzhou Research Base of China Agricultural University in Hebei Province. The summer maize Zhengdan958 was chosen as test cultivar. Five N treatments of N 0, 102, 145, 189 and 250 kg/hm² was designed and recorded as CK, OptN70%, OptN, OptN130%, ConN, respectively. The experiment has four replicates. The canopy images of summer maize at 6-leaf stage (V6) , 10-leaf stage (V10), tasseling stage (VT), blister stage (R2), filling stage (R4) were captured by using UAV-based visible spectrum technique. The canopy images were processed with Adobe Photoshop to extract color information of red (R), green (G), blue (B) and light (L) intensity. The correlation between the monitored 12 visible spectrum parameters calculated from R, G, B, L and plant N concentration, biomass and N uptake, the coefficient of variation (%) of visible spectrum parameters were researched to determine the best visible spectrum parameters for N status diagnosis of summer maize at different stages.

Results Field data showed that the redness intensity (R), greenness intensity (G), lightness intensity (L), normalized greenness intensity G/(R + G + B), normalized blueness intensity B/(R + G + B), the ratio of greenness and redness (G/R), the ratio of greenness and blueness (G/B), the ratio of greenness and lightness (G/L), the RGB vegetation index (RGBVI) had significant relationships with N concentration, N uptake and as well as plant biomass at all stages. Taking the coefficient of variation of visible spectrum parameters account, G/(R + G + B), G/L had higher correlation coefficient (r values 0.641–0.944) and smaller and stable coefficient of variation, ranged from 0.93% to 4.30%, compared with other parameters. So G/(R+G+B) and G/L can be the best visible spectrum parameters for N status diagnosis of summer maize at different stages.

Conclusions The UAV-based visible spectrum technique is usable in detecting N status dynamically and has the advantages of stable results, convenient and fast, high-efficiency and non-destructive. This study provides a scientific basis for promoting the application of UAV-based visible spectrum technique to diagnose crops N status dynamically on large scales.