Dynamic Monitoring Of Field Maize Growth Parameters Under Drought Stresses Based On UAV Remote Sensing

Maize is one of the major food crops in China.Affected by global warming,drought stress is seriously affecting the growth and yield of maize in arid and semi-arid regions of China.Agricultural drought is an important factor affecting maize production,and drought disaster is regional and has a certain degree of gradual,the occurrence of drought is not easy to be detected in time.Therefore,it is significant to innovate the way of monitoring and early warning of corn drought.UAV remote sensing has the advantages of fast acquisition speed,high resolution,low environmental impact,low cost and easy operation,which has great potential for application in crop growth dynamic monitoring,crop phenotype monitoring and disaster early warning research.In this study,we collected visible and multispectral images of maize at several key growth stages under different drought stresses based on UAV remote sensing,and simultaneously collected data of maize growth parameters on the ground to construct inversion models of different maize growth parameters,so as to estimate and obtain spatial and temporal distribution of maize phenotypic information in the field,to monitor the dynamics of maize growth parameters in the field under different drought stresses,to explore the response characteristics of maize phenotypic information under different drought stress,and to provide technical support for fine field management.The main research contents and conclusions of the paper are as follows.(1)A new method was proposed to extract phenotypic information of maize leaf color and plant height based on UAV visible image.Maize leaf color was obtained by color model transformation(RGB color model was transformed into HIS color model)and decision tree classification.Plant height was obtained by calculating the difference between Digital Surface Model(DSM)and Digital Terrain Model(DTM).It was feasible and accurate to extract maize canopy phenotypic information based on the above methods.(2)A new method for extracting maize canopy SPAD and photosynthetic parameters based on UAV multispectral images was proposed.Based on the canopy spectral reflectance information,combined with the maize canopy phenotypic parameters measured simultaneously on the ground,the correlation analysis between the vegetation indices and the ground collected data was conducted to screen the sensitive vegetation indices.Then,multiple stepwise regression,support vector machine and BP neural network models were used to establish remote sensing monitoring models for different canopy phenotypic parameters,and the optimal prediction model was selected to extract the canopy SPAD value and photosynthetic parameters of maize under different drought stresses at different growth stages.The results showed that for the two types of phenotypic information,the remote sensing monitoring model based on BP neural network had the strongest applicability to different growth stages and drought stress gradient,followed by support vector machine,and the multiple stepwise regression model had the worst applicability.(3)The effects of drought stress on the monitoring accuracy of maize canopy phenotypic information by remote sensing were revealed.With the intensification of drought stress,the prediction accuracy of models at different growth stages decreased significantly.However,the decrease of tasseling stage was smaller than that of other stages,and the inversion accuracy of the stage was better under different drought stresses.(4)Dynamic monitoring of maize phenotypic information changes at different growth stages.With the growth process of maize,the distribution of yellow canopy leaves and bare soil became less and less,while the distribution of yellow-green and green canopy leaves gradually increased.Maize plant height continued to grow and increased significantly at the stage of jointing and tasseling,but the growth rate decreased significantly at the later stage of maize growth.However,the growth rate decreased significantly in the late growth stage of maize,because the development of stem and leaf was the main development in the early growth stage of maize.The SPAD values of maize canopy increased gradually with the advance of growth stage,and chlorophyll content increased rapidly at the seedling-jointing stage,which was the stage of rapid development of maize leaves.Pn,Gs,Ci and Tr all increased firstly and then decreased during the whole growth stage,and all reached the maximum value at tasseling stage.(5)Dynamic monitoring of maize phenotypic information changes under different drought stress levels.There was no significant difference in the four phenotypic information between mild and moderate stress and CK,indicating that maize had certain adaptability and resistance to drought stress.Starting from moderate stress,drought stress significantly affected phenotypic information.With the aggravation of drought stress,the SPAD value of maize canopy tended to the same level in each growth stage,indicating that severe drought stress would lead to the obstruction or cessation of maize growth process,resulting in no significant fluctuation of chlorophyll level.Ci showed significant fluctuations compared with Pn,Gs and Tr at the beginning of drought,and had no significant difference under severe and extra severe stress,indicating that Ci was more sensitive to drought stress.(6)Response characteristics of maize phenotypic information under different drought stress.Under the same level of drought stress,physiological characteristics such as SPAD values and photosynthetic parameters of maize were significantly more correlated with soil drought than geometric characteristics such as leaf color and plant height,indicating that physiological characteristics of maize were more sensitive to soil drought.The leaf color of maize showed high correlation with drought stress only in the middle and late growth stage,indicating that the change of leaf color of maize under drought stress mainly concentrated in the middle and late growth stage.The plant height of maize were sensitive to different level of drought stress at different growth stages.The most sensitive level of drought stress at jointing stage was no stress,moderate stress at tasseling stage,and severe stress at maturity stage.With the development of maize growth stage,the most sensitive level of drought stress at plant height shifted from no stress to more severe stress.