Spectral Characteristics Of Corn Under Iron Toxicity Stress And Chlorophyll Concentration Inversion Experiment Research

Mining development may bring serious damage to the environment human’s life and production. Corn is one of the most widely distributed crops in the world as well as in iron mining area. However, iron toxicity for crop, which posed a great threat to food security, did not cause enough attention. Remote sensing could provide rapid, large-scale and non-destructive monitoring, and it is a powerful technology of environmental monitoring in the mining area. Therefore, spectral characteristics of corn under iron toxicity stress were studied by means of field and indoor experiments.First of all, the Fe content, chlorophyll, green peak and its location, red edge parameters of com organization were analyzed, and the spectral characteristics of leaf under iron toxicity stress were extracted. Secondly, according to the correlation of leaf spectral reflectance and chlorophyll concentration, sensitive spectral bands were selected to improve the spectral index. Finally, the improved spectral index was used to establish the model of chlorophyll inversion. The experimental results showed that, (1) When the Fe concentration in soil(or hydroponic solution) rose, the chlorophyll concentration rose first and then decreased. At low concentration (N1), Fe promoted the plant growth and chlorophyll concentration rose. At medium and high concentration (N2 and N3), Fe inhibited the plant growth and chlorophyll concentration decreased. (2) The green peak and its location, red edge parameters could reflect the change of corn leaf chlorophyll. At low concentration (N1), green peak reflectivity decreased, green peak position moved to blue direction (blue shift), red edge moved to red direction (red shift).(3)The correlation of improved spectral characteristics index mND705’, mSR705’, TCARI’and leaf chlorophyll concentration was increased. Among them, mND705’, the highest related index, was used for chlorophyll inversion by both statistic model and BP neural network. The result showed that the inversion error of BP neural network was 1.688SPAD and the inversion accuracy is higher than statistic model.Through the above research, the spectral characteristics were detected under iron toxicity stress. It may benefit monitoring iron pollution by use of hyperspectral remote sensing.