Cotton Canopy Information Recognition Based On Visible Light And Thermal Infrared Image Of UAV

In recent years,UAV have been widely used in agricultural monitoring because of their low price,flexible operation and high spatial and temporal resolution.This experiment used Airborne ordinary optical camera to collect cotton canopy images of different densities(15000,33000,51000,69000,87000,105000 plants/hm~2)under field production conditions.The stitched image is used to extract color information and construct a new color indexes.On the same day,the cotton growth characteristic parameters LAI,interception rate of PAR,and above-ground biomass were acquired.Using color indexes to establish the simulation equation of important characteristic parameters of cotton and verify it.The canopy temperature of cotton field under physiological stress after defoliating agent application was monitored by airborne thermal infrared imaging camera.At the same time,the upper leaves of cotton were monitored by temperature probe.We analyzed the change of cotton canopy temperature and upper leaf temperature under physiological stress.The test results are as follows:1.Using the unmanned aerial vehicle equipped with ordinary optical camera to extract cotton canopy information and make correlation analysis with cotton LAI,the results show:The exponent model of bgr is the best model for estimating LAI in two years(2017:R~2=0.6052017,RMSE=0.524;2018:R~2=0.5596,RMSE=0.469).Therefore,the cotton LAI in the whole fertility can be quickly estimated by using the color indexes bgr extracted from the digital image of the drone.2.Using the unmanned aerial vehicle equipped with ordinary optical camera to extract cotton canopy information and make correlation analysis with the above-ground biomass,the result show:The exponent model of 17 color indices is the best estimation of the above-ground biomass.The above-ground biomass that simulated by the exponent equation of bgr and the actual above-ground biomass are well correlated.However,the R~2 of the bgr color indexes is small(2017:R~2=0.3516,2018:R~2=0.485)and the RMSE is large(2017:RMSE=3396.54,2018:RMSE=1529.23),so the estimation of above-ground biomass is not accurate good..3.Using the unmanned aerial vehicle equipped with ordinary optical camera to extract cotton canopy information and make correlation analysis with the interception rate of PAR,the result show:The quadratic model of 17 color indices is the best estimation of the interception rate.of PAR.The interception rate of PAR that simulated by the quadratic equation of b and the actual interception rate of PAR are well correlated(2017:R~2=0.7412,RMSE=0.125,2018:R~2=0.8674,RMSE=0.059).Therefore,we can use the color indexes b that extracted from the digital image of the UAV to monitor the interception rate of PAR.4.Using the unmanned aerial vehicle equipped with thermal infrared camera to monitor the cotton fields treated with defoliants continuously,the result show:The average cotton canopy temperature of the treated group was significantly higher than thatthe control group at noon with the highest solar radiation intensity.Therefore,noon is the best time to monitor the canopy temperature of cotton field under physiological adverse conditions by airborne thermal infrared imaging.With the increase of the days after using defoliate,the temperature difference of cotton canopy in the noon treatment group and the control group gradually increased and then stabilized.At this time,the cotton leaves are basically detached,so the defoliation effect can be monitored by using airborne thermal infrared imaging technology.In summary,the color indexes extracted from the cotton canopy image obtained by the UAV's visible light camera can be used to estimate the cotton LAI and the interception rate of PAR.The airborne thermal infrared imaging can well monitor the canopy temperature of cotton under physiological stress conditions.It lays the foundation for precise farmland management.