Research On Rice Cadmium-Stress Monitoring Based On Fusion Mechanism Model And Measured Data

If crops are grown in a soil environment with excessive heavy metals for a long time,the crops will be subjected to stress,and their physiological and biochemical parameters will change,resulting in changes in the spectral response characteristics of the crops.Because remote sensing technology can accurately,real-time,continuously and dynamically acquire the spectral response information of crops to environmental stress,the use of remote sensing technology to monitor the status of crop heavy metal pollution has gradually replaced traditional ground survey methods.At present,the remote sensing monitoring of crop heavy metal stress is mostly using statistical models constructed empirical/semi-empirical model based on vegetation spectral characteristics or spectral indices,but the models lack robustness and universality.Because the radiative transfer models of the physical process can overcome the lack of robustness and universality of the empirical / semi-empirical model,it provides new research ideas for remote sensing monitoring crops of heavy metal pollution in largescale.In this paper,the rice planting area in Zhuzhou area of Hunan Province,which is seriously polluted by farmland soil cadmium(Cd),is selected as the study area.The physiological and biochemical parameters such as soil Cd content,rice chlorophyll content,leaf area index and leaf water content are collected.The Sentienl-2A image data in 2017 and 2018 are collected to explore the effect of soil Cd pollution on the physiological and biochemical functions of rice and its spectral characteristics.The main research work and conclusions are as follows:(1)Improve the traditional canopy spectral model PROSAIL model,determine the optimal value of leaf structure parameter N,introduce the absorption coefficient of cadmium,and construct the canopy spectral simulation model Cd-PROSAIL model under Cd stress.Compared with the original PROSAIL model,the spectral reflectance of rice under Cd pollution simulated by the new model is improved.The improved CdPROSAIL model has higher accuracy in the spectral simulation results and is closer to the measured spectral.(2)Based on the sensitivity analysis to analyze the impact of Cd pollution stress on the physiological and biochemical parameters of rice on the spectral reflectance,combined with the band correlation analysis method,a sensitive spectral index(NDVI_Cd)to Cd stress response is constructed.NDVI_Cd can better distinguish the different stress levels of rice than the existing remote sensing monitoring indexes of heavy metal stress(such as NDVI,DVI,etc.),especially under heavy stress.(3)According to the look-up table method,the improved Cd-PROSAIL model is used to establish the Cd stress rice spectral database,and the pixel-by-pixel matching is performed on the Sentenl-2A remote sensing image to obtain the regional distribution of Cd in rice in the experimental area.