Quantitative Inversion Of Apple Canopy Parameters Based On HJ-1A-HSI Data And PROSAIL Model

The chlorophyll content of crops is an important indicator reflecting its photosynthetic capacity and its own growth and development.The Leaf Area Index(LAI)is an important basis for estimating the surface coverage of green plants,predicting crop growth trends and yields.Traditional methods for obtaining vegetation chlorophyll and leaf area index are generally time-consuming and laborious,and it is difficult to meet the needs of large-area monitoring.The method of estimating the biochemical parameters of a wide range of vegetation can be realized by using hyperspectral data images combined with the physical radiation transmission model,which greatly improves the accuracy of spectral quantitative inversion of vegetation parameters,and predicts crop growth,fertilization regulation and yield evaluation.It is of great significance.In this paper,the Qixia City of Shandong Province was used as the research area,and the PROSAIL radiation transmission model was used.The HJ-1A-HSI data combined with the measured data of the sampling points were used to invert the chlorophyll content and leaf area index of apple canopy for monitoring.Apple leaf chlorophyll content and LAI provide a way to monitor fruit tree yield.The main findings are as follows:(1)The canopy reflectivity inversion and accuracy evaluation of apple number based on HJ-1A-HSI data were performed.Data reading,absolute radiance value conversion,strip removal,atmospheric correction,geometric fine correction and study area cropping were performed on the image data,and the vegetation spectrum curves of 34 samples were extracted.The measured sample reflectance data is resampled to the image,and the image data before and after the pre-processing is simultaneously compared and fitted.The results show that the spectral curve after pretreatment is basically consistent with the measured spectral curve of the canopy.The coefficient of determination R~2 is increased from 0.005 to0.9714,and RE%is reduced from 10.6169 to 0.6202.Through the reflectivity inversion process of HJ-HSI data,the data quality of the image is greatly improved,and the results are consistent with the ground canopy data.It can be used to quantify the leaf area index and chlorophyll content of apple canopy.(2)The simulated canopy reflectivity of apple canopy based on PROSAIL model was constructed.During the simulation,the input values of leaf parameter index,chlorophyll content,carotenoid content,dry matter content and equivalent water thickness of the model parameters were taken as the experimental average.The input values of the solar elevation angle,the observed zenith angle and the azimuth parameters are obtained directly from the HSI image data.Psoil,ALA,Cbrown,hspot,and other parameters are based on previous studies to take constants as model input values.The measured canopy reflectance data is fitted to the model simulated reflectance data.The results show that the determination coefficient R~2is 0.9944 and the RE%is 0.1845.The fitting accuracy of the two is high.PROSAIL can be used for the inversion of apple canopy parameters.(3)The parameter sensitivity analysis of the PROSAIL model was performed.The parameter sensitivity analysis uses a qualitative analysis method for the full band of the PROSAIL model,the leaf area index(LAI),the chlorophyll content(Cab),the carotenoid content(Car),the equivalent water thickness(Cw),and the dryness of the PROSAIL model.Sensitivity analysis was performed on six parameters of material content(Cm)and leaf structure parameters(N).The results show that in the visible light range,the sensitivity of each parameter is from high to low,LAI>Cab>N>Cm>Cw>Car;in the infrared band,LAI,N,Cm and Cw are sensitive,and Car and Cab sensitivity is higher.Low,almost no.Combined with the HSI data band range(450nm~950nm),the sensitive parameters are determined as LAI,Cab,Cm,and N.(4)The apple canopy LAI and chlorophyll inversion based on PROSAIL model were completed.Using the cost function minimization method to determine the blade structure parameter N is 1.5;the leaf area index LAI is 1.5~4.5,the step length is 0.1;the chlorophyll content Cab range is 55~105,and the step size is 2(?g/cm~2);The dry matter content Cm ranges from 0.016 to 0.04,and the step size is 0.002(g/cm~2);the remaining parameters are measured by the measured mean or constant value.Enter the parameter values or range of values into the model and establish a LAI-Cab-Apple Canopy Reflectance Lookup Table by setting the cycle.The parameters LAI and chlorophyll content were inverted by combining the pre-treated HSI data.The results show that the L2 model inversion value and the measured value of the measured coefficient R~2 are 0.8556,the root mean square error RMSE is 0.2407,and the fitting effect is better.And the RPD is 2.2740,and the value is greater than2,indicating that the model has good inversion ability for LAI.The decision coefficient R~2 of the model inversion and measured value of Cab is 0.9159,the value is ideal,the root mean square error RMSE is 2.9947,the value is too large,but the RMSE value has been reached in combination with the range of chlorophyll content and its step size.The ideal value,the inversion value and the measured value fit well.And the RPD of the two is 3.0290,the value is greater than 2,indicating that the model has better inversion ability for Cab.