Diagnosis of heavy metal cross contamination in leaf of rice based on hyperspectral image: a greenhouse experiment

Abstract

The objective of this study was to explore the feasibility for rapid diagnosis of heavy metal Pb and Cd cross contamination in agricultural soil with hyperspectral image of rice plants. Several data pretreatment methods were applied to improve the diagnosis accuracy. The best diagnosis accuracy of SVM model were all above 75% for any none, low, medium and high concentration stress based on the spectral band of second derivative. To the best of our knowledge, this is the first study identifying Pb and Cd cross contamination by using hyperspectral images. The results indicates that it is feasible to diagnose the Pb and Cd cross contamination in agricultural soils from observation of rice plants using hyperspectral image.