| Tea is an important economic crop in China,and recent surveys have shown that the production and economic benefits of tea in China rank first in the world.The content of chlorophyll a and b in tea is an important evaluation index of the physiological and ecological characteristics of tea trees.Measuring the content of chlorophyll a and b in tea trees can help understand the growth state of tea trees,predict tea production,evaluate the drought and cold resistance of tea,which is of great significance for production life.At present,the main method to measure the chlorophyll a and b content of tea leaves is through chemically destructive detection,and the remote sensing inversion method suitable for large-scale tea plantations can only detect the total amount of chlorophyll.Therefore,the remote sensing inversion method needs to be improved so that it can invert chlorophyll a and b contents separately.In order to obtain more accurate hyperspectral data,radiometric correction is needed,and a portable and rapid non-destructive instrument for detecting chlorophyll a and b in leaves is required to verify the accuracy of the improved remote sensing inversion method for inversely detecting the content of chlorophyll a and b in tea tree canopy.Hyperspectral remote sensing detection is an effective nondestructive detection method,and hyperspectral reflectance data is commonly used for agricultural crop status detection,prediction of crop yield and phytochrome inversion,etc.Therefore,this study uses hyperspectral data and pseudo-absorption overlapping band separation method to achieve sensitive band and suitable bandwidth acquisition for detecting chlorophyll a and b,and develops a portable instrument for chlorophyll a and b content detection;in order to obtain more accurate spectral information a radiation correction model is used to calibrate hyperspectral data in the complex environment of tea plantations;based on the above study,the improved PROSAIL model was used to invert the chlorophyll a and b content of the tea tree canopy,and the inversion accuracy was verified using the portable instrument.The following studies and results were conducted based on the above methods:(1)Development of plant leaf chlorophyll a and b content detection instruments:(1)analyze and compare their inversion accuracy by two hyperspectral methods,spectral index and pseudo-absorption coefficient,and determine the best sensitive bands for chlorophyll a and b detection as 650 nm,700 nm and 900 nm;(2)The spectral consistency at four different resolutions was verified by comparison of spectral downscaling methods,and the results showed that the use of low-resolution spectral information of 16 nm in its sensitive band for chlorophyll a and b inversion could replace the high-resolution(3nm)spectral information,and the 16 nm bandwidth without any reduction in accuracy was chosen as the bandwidth of the sensitive band of the chlorophyll a and b detection instrument under the consideration of cost;(3)Based on the above study,a portable chlorophyll a and b detection instrument was developed based on the photoelectric detection technology using LED lights with sensitive bands(650 nm,700 nm,900 nm)in the center band and 16 nm bandwidth filters as light sources,and a photosensitive sensor to obtain the reflectance of sensitive bands of leaves.The final RMSE of chlorophyll a and b content detection accuracy for the instrument was 2.2μg/cm~2and 1.1μg/cm~2,with the accuracy of93.8%and 92.3%,respectively,indicating that the leaf chlorophyll a and b instrument can effectively detect chlorophyll a and b content,which provides a tool for subsequent rapid detection of chlorophyll a and b content in tea plantations.(2)In order to more accurately obtain the hyperspectra data of tea tree canopy that can be used to invert chlorophyll a and b,radiation correction model research was conducted:The tea plantation canopy hyperspectral data were corrected using the cosine correction,C correction,SCS correction,SCS+C correction model applicable to the Lambertian surface and the BRDF radiation correction model applicable to the non-Lambertian surface,where the BRDF correction model for the non-Lambertian surface was subdivided into the considered case and the case without topography.Then we compare the correction effect of the main characteristic bands(670 nm,620 nm,550 nm,780 nm)of chlorophyll inversion using the above radiation correction models and find that the Lambertian correction is not good and the correction effect of the non-Lambertian case without considering the topography is better than the effect of considering the topography.The error ratios of the four characteristic bands were reduced from 9.14%to 4.97%,6.59%to 3.60%,7.17%to2.64%,and 2.21%to 1.78%after the correction.The errors in the spectral reflectance curves of tea tree canopies were significantly reduced and the color of hyperspectral pseudo-color images of tea plantations were more uniform after using the BRDF correction model without considering topography to correct the full spectrum of tea plantations,indicating that the correction model is effective in providing more accurate spectral data for subsequent studies.(3)Inversion of tea tree canopy chlorophyll a and b content based on a modified PROSAIL model:(1)An improved PROSAIL model capable of inverting the canopy chlorophyll a and b content was established by modifying the inner chlorophyll absorption coefficients in the PROSAIL model to the absorption coefficients of chlorophyll a and b respectively according to the latest PROSPECT-MP model in the400-800 nm range,and adjusting the mean refraction coefficient and carotenoid absorption coefficient by adding the baseline absorption coefficients;(2)Local sensitivity and global sensitivity analysis of inputs in the modified PROSAIL model Sensitivity from largest to smallest was chlorophyll b(Cb)>chlorophyll a(Ca)>carotenoids(Car)>leaf structural parameters(N)>leaf area index(LAI)>dry matter content(Cm)>equivalent water thickness(Cw),and determination of parameter Cm,Cw and N values to reduce data dimensionality;(3)Comparison of the inverse performance of tea tree canopy chlorophyll a and b contents based on the actual hyperspectral data after radiation correction and the simulated spectral data generated by the improved PROSAIL model and comparison with the mean values of tea leaf chlorophyll a and b contents measured by the leaf chlorophyll a and b detection instrument was verified in which the RMSE were 7.10μg/cm~2and 3.53μg/cm~2,with the accuracy of 88.02%and 81.35%respectively.This indicates that the improved PROSAIL can effectively invert the chlorophyll a and b contents of tea tree canopies.The above study can realize the detection of chlorophyll a and b content of tea trees in a large area and wide range of tea gardens,and the spectral data of tea gardens at different times and locations can be collected for database establishment,so that the changes of chlorophyll a and b content of tea trees in tea gardens can be obtained conveniently for a long time,thus the growth and physiological condition of tea trees can be assessed quickly,timely and accurately,which is conducive to promoting the management and improvement of tea quality and yield in tea gardens. |
PDF Full Text Request