Retrieval Of Vegetation Parameters At Canopy And Leaf Level Using Hyperspectral Remote Sensing

Vegetation is one of the most important components of the ecosystem,and dynamic change of vegetation cover can reflect the ecological stability and vulnerability.Physiological and biochemical parameters of vegetation are closely related to the material and energy exchange process.The accurate retrieval of vegetation has attracted the attention of scientists in the field of quantitative remote sensing for decades.However,there are still a lot of difficulties and uncertainties in vegetation parameters retrieval to date.In this study,we focus on three typical problems in current research of vegetation parameters retrieval: 1)the vegetation canopy BRDF(bi-directional reflectance distribution function)effect on the retrieval of vegetation parameters;2)the difficulties of retrieval of vegetation parameter at leaf level;3)the difficulty of separating the contribution of different vegetation components to the reflectance spectrum.To deal with those problems,we carried out the research on three aspects: 1)developed a new BRDF-resistant vegetation index for improving the estimation of leaf area index;2)retrieval of chlorophyll content at both canopy level and leaf level by remote sensing;3)carried out a new method for the retrieval of vegetation parameters based on Principal Component Analysis(PCA).The main conclusions are:(1)Numerous vegetation indices have been developed to estimate the LAI(Leaf Area Index).However,because of the effects of the bi-directional reflectance distribution function(BRDF),most of these vegetation indices are also sensitive to the effect of BRDF.We found the similarity of the BRDF curve shape in the solar principal plane between the green and the near-infrared(NIR)bands,and between the blue and red bands.The consistency of the shape of the BRDF across different bands was employed to develop a new BRDF-resistant vegetation index(BRVI)for estimating the LAI.The results showed that BRVI had the lowest diversity ratio(DR),which ranged from 1.01 to 1.05—the DR values for the simple ratio vegetation index(SR)is as high as 3.20.The potential of the proposed BRVI for estimation of the LAI was evaluated using both simulated data and in-situ measurements and also compared to other popular vegetation indices.The results showed that the influence of the BRDF on the BRVI was the weakest and the BRVI performs well for the LAI retrieval,with a coefficient of determination(R2)of 0.84 and an RMSE of 0.83 for the field data and with an R2 of 0.97 and an RMSE of 0.25 for the simulated data.Therefore,it was concluded that the new BRVI is resistant to BRDF effect and is also promising for use in estimating the LAI.(2)The vegetation spectral indices have good correlation to the canopy chlorophyll density,but the applicability of different vegetation index in the retrieval of chlorophyll content by remote sensing was different.By comparing the correlation between the 27 vegetation indices and canopy chlorophyll density,the coefficient of determination and retrieval error of retrieval model of canopy chlorophyll density based on vegetation index,we found that the MERIS Terrestrial Chlorophyll Index(MTCI)which contains the red-edge band information is the most optimal vegetation index for retrieving canopy chlorophyll density.We established retrieval model of canopy chlorophyll density based on MTCI.The model is suitable for retrieving canopy chlorophyll density in different growth stages and retrieving canopy chlorophyll density of different crop types.(3)The retrieval accuracy of canopy chlorophyll content is relatively good,but because of the strong interference of LAI,the retrieval of leaf chlorophyll content by remote sensing is still difficult.Based on the prior knowledge of vertical distribution of chlorophyll content in winter wheat canopy,an algorithm for estimating the canopy top leaf chlorophyll content of from canopy spectra was derived.The results based on insitu measurements showed that the scatters of simulated values versus measured values of the canopy top leaf chlorophyll content were distributed near the 1:1 line with RMSE of 0.014 mg/cm2.We could conclude that,with the vertical distribution of chlorophyll content as prior knowledge,the scale transforming model for chlorophyll content from canopy level to leaf level can be established.(4)The vegetation index method based on the combination of two or more bands has been widely used in retrieving vegetation parameters.However,the spectral information used in this way is limited.We explored a new method of retrieving vegetation parameters based on PCA.The principal components were extracted from reflectance spectra from simulated data,then the first several principal components were used to reconstruct the measured reflectance spectra.We found that after the spectral reconstruction,the weight coefficients of a specific principal component are well correlated with some specific vegetation parameters.Based on this theory,retrieval models of vegetation parameters could be established.The retrieval accuracy of chlorophyll content and equivalent water thickness is superior to the model based on traditional vegetation index.In addition,the retrieval accuracy of principal component analysis with the sensitive bands to vegetation parameters,is better than that of principal component analysis with the full spectrum.Principal component analysis with 400-800 nm bands could be used to retrieve the leaf chlorophyll density;and combination 400-800 nm and 900-2500 nm bands could be used to retrieve leaf chlorophyll concentration;and principal component analysis with 900-2500 nm bands could be used to retrieve leaf water content.The models established based on both full spectrum and sensitive spectrum could not retrieve dry matter content well.The main innovative contributions of this paper are as follows:(1)A new vegetation index,BRVI,which is BRDF-resistant and sensitive to LAI is developed based on the similarity of BRDF curves of red/blue band and NIR/green band.The new developed BRVI is efficient to improve the accuracy of LAI retrieval under varying imaging geometry conditions.(2)A scale transforming model for chlorophyll content from canopy level to leaf level were established based on the prior knowledge of the vertical distribution of chlorophyll content within the canopy,and a new algorithm for the retrieval of chlorophyll content in the top leaves is proposed based on the scale transforming model.(3)A new method for vegetation parameters retrieval based on the principal component analysis is proposed.The results indicate that there are strong relations between the weight coefficient of specific principal component or combination of specific principal components used for the reconstruction of measured reflectance spectrum and specific vegetation parameters.Compared with the method based on vegetation indices,the new PCA-based method has advantages in making use of more spectral information in the retrieval of vegetation parameters.