Monitoring Of Physiological And Ecological Phenotypic Indexes Of Foxtail Millet Based On Unmanned Aerial Vehicle Multispectrum

This paper is in order to study the vegetation index inversion of millet leaf area index,chlorophyll content,and other physiological and ecological phenotypic traits.During the experiments,we observed the millet planted in Dong Yang Experimental Base of Shanxi Agricultural University and used UAV equipped with multispectral cameras to collect multispectral images of millet at several stages of growth and development.And the vegetation index was extracted from it.We also collected the measured ground data at the same time.Various models for predicting the physiological and ecological indicators of millet were established,tested and verified.Under the condition of sufficient water supply treatment,the determination coefficient R2 of the LAI prediction model constructed based on NDGI was 0.79.The RMSE was 0.09 and the stability was better.The MAE was 0.12 and the accuracy was higher.The RMSSE was0.17 and the fit was better.Under drought stress conditions,the determination coefficient R2 of the LAI prediction model constructed based on NDVI was 0.81.The RMSE was 0.42 and the stability was better.The MAE was 0.51 and the accuracy was higher.The RMSSE was 0.18.and the fit was also better.Under the condition of sufficient water supply treatment,the determination coefficient R2 of the chlorophyll content prediction model based on NDVI was 0.81.The RMSE was 1.31 and the stability was better.The MAE was 1.06 and the accuracy was higher.The RMSSE was 0.16 and the fit was also better.Under drought stress conditions,the determination coefficient R2 of the chlorophyll content prediction model constructed based on NDVI was 0.71.The RMSE was 6.71,and the stability was poor.The MAE was 5.96 and the accuracy was low.But the RMSE was 0.76,the fit was good.Under the conditions of sufficient water supply treatment,using the quadratic function method,the determination coefficient R2 of the transpiration rate prediction model constructed based on the Blue band was 0.87.The RMSE was 0.00039 and the stability was better.The MAE was 0.27605 and the accuracy was higher.The RMSSE was 0.16642 and the fit was better.Under drought stress conditions,using the quadratic function method,the determination coefficient R2 of the transpiration rate prediction model constructed based on the Blue band was 0.83.The RMSE was 0.00032 and the stability was better.The MAE was 0.37369 and the accuracy was higher.The RMSSE was 0.17584 and the fit was better.Under the conditions of sufficient water treatment,using the quadratic function method,the determination coefficient of the net photosynthetic rate prediction model based on PIRI was0.58.The RMSE was 1.84889 and the stability was better.The MAE was 0.17246 and the accuracy was higher.The RMSSE was 0.27038 and the fit was better.Under drought stress conditions,all vegetation indices and net photosynthetic rates were insignificant at a significant level of 0.05,so it was impossible to construct a prediction model for millet net photosynthetic rates under drought stress conditions.