Determination of Corn Nutrient Status under N&K Stressed Condition Using Hyperspectral Analysis

Variable fertilization for crops, like corn, depends on monitoring nutrition condition. Thus, hyperspectral reflectance was used to predict chlorophyll and total nitrogen content under different N,K treatments during corn growth. The area of the experimental field was divided into 3 strips with different nitrogen treatment (N1:0 kg/ha-low, N2:314 kg/ha-normal, and N3:653 kg/ha-high). In each strip, there were 3 repetitions of 3 potassium treatments(K1:0 kg/ha-low, K2:214 kg/ha-normal, and K3:500 kg/ha-high). The results show that growth stages happened in advance under low nitrogen treatment and reflectance intensity, total N, and chlorophyll content are all largest under N2K1 in shooting stage. GNDVI(R2=0.88, RMSE=0.08) performs well for chlorophyll prediction under N3K2. In addition, MLR has the potential of determination of chlorophyll(R2=0.94, RMSE=0.02) and total N(R2=0.97, RMSE=0.09) in trumpet stage, as well as PLSR for chlorophyll(R2=0.99, SEC=0.01;SEP=0.09) and total N (R2=0.96, SEC=0.11;SEP=0.47) in trumpet stage.