| Chlorophyll and carotenoid are two major photosynthetic pigments in crop leaves,which are often used to indicate their photosynthetic capacity and the growth status of crops.Timely and accurate monitoring of photosynthetic pigment content of crops can not only provide decision-making information for agricultural managers,but also realize agricultural water and fertilizer management,so as to improve the utilization rate of water and fertilizer,which has extremely important significance for the development of sustainable modern agriculture.The development and wide application of hyperspectral technology make it possible to monitor crop growth in real time and accurately.However,compared with the traditional vertical observation and remote sensing method,multi-angle observation is helpful to obtain the three-dimensional structure characteristic information of ground targets,so the information of ground objects can be retrieved more accurately.This study was based on winter wheat as the research object and analyzed the correlation between canopy spectra under different observation angles and the vegetation index and winter wheat photosynthetic pigment content based on the multiple perspectives of winter wheat canopy spectral multi-angle reflection of canopy spectral features,combined with Partial Least Squares Regression(PLSR)to select the optimal observation point.This study finally used Multiple Linear Regression(MLR)provision,Partial Least Squares Regression(PLSR)and Support Vector Machine Regression(SVR)to simulate the winter wheat photosynthetic pigment content inversion,and used the R2,RMSE and RPD to comprehensively evaluate and chose the best prediction model and the best observation angle of photosynthetic pigment content in winter wheat canopy.The conclusions of this study are as follows:(1)The decrease of chlorophyll content and carotenoid content would lead to the increase of canopy reflectance of winter wheat,and the content levels of both were inversely proportional to spectral reflectance.Among all the observation angles,the canopy spectral reflectance observed at 30° backscattering was the highest,because the backscattering direction was consistent with the direction of solar incident,so that the spectrometer sensor could obtain more reflection information.(2)Different observation angles had different effects on the correlation between canopy spectrum,vegetation index and chlorophyll and carotenoid content.In all the observation angles,the correlation between 30° backscatter canopy spectrum and vegetation index and photosynthetic pigment content of winter wheat was the highest.Among them,the correlation between vegetation index TVI and chlorophyll a,b and carotenoids in the canopy reached the highest level(0.617,0.556 and 0.602,respectively),which all reached a significant correlation at the level of 0.01.(3)Different observation angles of chlorophyll and carotenoid monitoring model of the effect also has great influence.For chlorophyll,the best model is based on the backward scattering field of vegetation index of 30° canopy chlorophyll a Multiple Linear Regression model(R2 = 0.731,RMSE = 3.413,RPD = 1.545).For the chlorophyll b,the best model was based on the backward scattering field of vegetation index of 60°canopy chlorophyll b Support Vector Machine Regression(SVR)model(R2 = 0.623,RMSE = 1.286,RPD= 1.527).For carotenoids,the best model was the Partial Least Squares Regression model of canopy carotenoids based on full spectrum backscattering of 60°(R2=0.510,RMSE=0.758,RPD=1.423).The chlorophyll and carotenoid monitoring model established in this study has better predictive ability,and could better achieve the accurate quantitative estimation of chlorophyll and carotenoid compared with the traditional observation perspective.Therefore,it was feasible to use multi-angle hyperspectral remote sensing technology to improve the quantitative monitoring accuracy of photosynthetic pigment in winter wheat canopy. |
PDF Full Text Request