Hyperspectral Quantitative Remote Sensing Inversion Model And Regieme Of Multiple Pigments At Leaf Scale Based On PROSPECT-PLUS Model

The information characteristics of leaf pigments are an important indicator for plant physi-ological and ecological status. And absorption spectral characteristics of these pigments are lo-cated in the higher energe spectra regions where sunlight reachs earth surface. With the devel-opment of remote sensing technology, hyperspectral remote sensing has been regarded as the effective method for the detection or monitoring of the vegetational pigment information. So, the vegetational physiological and ecological information is indirectly acquired through the de-tection or monitoring for the change on leaf pigment information in 400-800 nm based on the hyperspectral remote sensing technology.Leaf spectra is controlled by the effect factors of leaf optical properties, and the description of leaf optical properties through quantifying these effect factors based on the leaf optical radia-tive transfer model is an effective method for the detection of leaf information characteristics. Therefore, PROSPECT leaf optical radiative transfer model in our study, which has been widely applied to the vegetational optical radiative transfer, was employed for the further quantification of the optical response factor of leaf physiology and ecology and leaf surface geometrical char-acteristic factor, and then set up the leaf pigment optical radiative transfer model (PRO-SPECT-SPPP, PROSPECT-SGED, PROSPECT-PLUS) for leaf spectra modeling and pigment concentration inversion.In the absorption coefficient function from PROSPECT model, there are the band overlap-ping between the pigment specific absorption coefficients and the higher linear correlationship between the pigment contents, which blocks the spectra modeling of leaf with multiple pig-ments and the corresponding multiple pigments concentration inversion. In our study, we em-ployed the G-L function from the spectra separation peak to describe the corresponding pig-ments specific absortion coefficients in the leaf optical radiative transfer procedure, and modi-fied and developed the function item that can reflect the pigment information (Chla, Chlb, Cars, and Ants specific absorption coefficients function item and the corresponding pigment content function item), which could transform the parallel parameters from the leaf absorption function into the non-paralled parameters from the G-L function for the each absorption peak of those pigment specific absorption coefficients in the leaf radiative transfer course. And then, em-ployed the corresponding pure pigment absorption spectra characteristics in the organic solution to gain the peak position for each absorption peak of each specific absorption coefficients and introduce a red-shif parameter to revise each peak position relationship between in organic solu-tion and in vivo leaf. Those treatments on the leaf absorption coefficient function in leaf optical radiative transfer aimed to gain the separated Chla, Chlb, Cars and Ants specific absortion coef- ficients in vivo leaf and acquired a leaf pigment optical radiative transfer model (PRO-SPECT-SPPP) in 400-800 nm incorporating leaf specific photosynthetic (Chla, Chlb, Cars) and photo-protective (Ants) pigments information characteristics.For leaf surface geometry characteristics in PROSPECT model, a constant idea maximal incident angle (59°or 40°) was employed for the defination of the solid angle of incident light that could describe the quantificational relationship between the incident light and leaf surface geometrical characteristics, which limited the accuracy of leaf pigment concentration inversion from the different leaves with the surface geometrical characteristics and blocked the pigment inversion using the hyperspectral remote sensing data with the non-nadir illuminating direction. To improve those pigment inversion abilities for PROSPECT, DHRFspec model was em-ployed for the quantificational description of the leaf surface interface reflection in the frame of leaf optical radiative transfer PROSPECT model, in which the introduced leaf surface rough-ness parameter could describe leaf surface geometrical characteristics and the input variable for the introduced incident light zenith angle could depict the illuminating light direction. Thus, the quantificational description for the SHADOWED and MASKED phenomenon that incident light impinges on the leaf surface rough facet of the "V" shape at microscale from non-nadir illuminating direction could be performed using the two introduced paramters in PROSECT optical radiative transfer frame. And then in 500-800 nm we could acquire a PRO-SPECT-SGED model with an ability for the quantificational describing leaf surface geometrical characteristics and a Chls concentration inversion function using the data from the non-nadir illuminating light direction, which is a coupling between PROSPECT and BRDF.Furthermore, in 500-800 nm region of the PROSPECT leaf optical radiative transfer frame, we further describled leaf physiological and ecological response pigments and leaf surface ge-ometrical characteristics of leaf optical property effect factors simultaneously, through employ-ing G-L function from the separating peak technology of the absorption spectra to determine the pigment specific absorption coefficients with the band overlapping characteristics, and employ-ing DHRFspec model from BRDF to quantificationally express the leaf surface interface reflec-tion radiance. And then set up a leaf pigment optical radiative transfer model (PRO-SPECT-PLUS) in 500-800 nm, which could provide an ability for the pigment concentration inversion incorporating Chla and Chlb at the non-nadir illuminating light direction and is a sup-plement for the coupling between PROSPECT and BRDF.And the new ZHELOP dataset and seclect LOPEX93 dataset and the new NNDHRF da-taset were collected to determine the model parameters and model function test for the corre-sponding ROSPECT-SPPP and PROSPECT-SGED and PROSPECT-PLUS, respectively. And the comparsions for the determined corresponding model parameters and the results of model function test were performed between those new built leaf pigment radiative transfer model and the former PROSPECT models. The results is mainly showed:(1) the multiple leaf pigments with the band overlopping characteristics are successfully determined by PROSPECT-SPPP model, for example, Chla, Chlb, Cars and Ants specific absorption coeffecients are successfully in the ZHELOP dataset. For the application of the PROSPECT-SPPP model, it can simulate the optical properities of leaf with multiple pigments, and can invert those pigments concentration using leaf reflection and transmission spectra. Comparing with PROSPECT-5 model, this model can improve the modeling accuracy of leaf reflectance and transmittance, and broaden the as-sessed pigment specis in vivo leaf. (2) PROSPECT-SGED model can invert the Chls concentra-tion using leaf reflection spectra with the light source form the arbitrary incident angles. Com-paring with PROSPECT-4 model, this model can improve the assessed accuracy of Chls con-centration using leaf reflection spectra with a light from the nadir direction (3) PRO-SPECT-PLUS model can simulate the optical properties of leaf with Chla, Chlb presents and the nadir incident light in 500-800 nm, and can invert Chla, Chlb concentration using the leaf reflectance and transmittance in 500-800 nm with an incident light from the nadir direction, and can also can invert Chla, Chlb concentration for the incident light from the non-nadir direction. Comparing with PROSPECT-SPPP model, PROSPECT-PLUS model can improve the assessed accuracy of Chla, Chlb concentration for the incident light from the nadir direction,that the determined model parameters for ROSPECT-SPPP and PROSPECT-SGED and PROSPECT-PLUS were reasonable; that the evaluation function values based on RMSE and BIAS and SE appeared to be credible for the spectra modeling of those new built models with a illuminating light at nadir direction; that to the inversion of the correspongding pigment con-centration with the illuminating light at nadir and non-nadir direction; the evaluation function values from the RMSE and BIAS and SE and VC function showed to be encouraging that these new leaf pigment optical radiative transfer models are applied to the required pigments inver-sion.In conclusion, in our study the science problems, which are the separation of the band overlapping characteristics between all kinds of pigment specific absorption coefficients and are the quantificational description for the SHADOWING and MASKED phenomenon that incident light impinges on the leaf surface rough facet of the "V" shape at microscale from non-nadir illuminating direction in PROSPECT leaf optical radiative transfer frame, were solved by em-ploying the G-L function from the separating peak technology for absorption spectra and em-ploying the DHRFspec model from BRDF. And then we set up the leaf pigment optical radia- tive transfer models (PROSPECT-SPPP and PROSPECT-SGED and PROSPECT-PLUS), those new built model can improve and broaden the PROSPECT spectra modeling of leaf incorporat-ing the subdividable multiple pigments (Chla, Chlb, Cars and Ants) and the corresponding pig-ment concentration inversion at non-nadir illuminating direction, which is a supplement for PROSPECT model and a progress for the coupling between PROSPECT and BRDF.