Multi-spectral Spatial Distribution Model Of Leaf Area Index And High Resolution Remote Sensing Image Retrieval

The leaf area index(LAI)inversion is important to the estimation of terrestrial vegetation biomass and the monitoring of ecosystem status.The LAI inversion plays an important role in the investigation of crops and desert vegetation in arid and semi-arid regions of northwestern China.The appearance of high-resolution remote sensing images provides a possibility for high-precision leaf area index retrieval.In this study,the input parameters of the PROSAIL model were determined using field observation data.The maize,rice,desert scrub,and grassland in Zhongwei Oasis of Ningxia were selected as the research object,and the relationship between these vegetation reflectance and LAI was discussed.Different red/near-wavelength reflectivity-LAI look-up tables were used to retrieve the small-area leaf area index for the WorldView-3 remote sensing image,providing a methodological reference for retrieval of leaf area index of highresolution remote sensing images.The main conclusions are:(1)The input parameter calibration is the first step in the LAI inversion based on the PROSAIL model.Some of the input parameters of the PROSAIL model cannot be obtained or are difficult to obtain.Because these parameters have an influence on the simulation results,the parameter determination process is more objective in determining the parameter values,which can effectively improve the inversion accuracy of the leaf area index.(2)When building a simulation data set,input schemes can be formulated for different parameters of the model.The input parameters of the PROSAIL model are numerous,and all involved in constructing the simulation data set will lead to an excessively large dimension of the simulation data set,which will reduce the computational efficiency,and the response of the red,near-infrared band reflectance to the change of the leaf area index will decrease as the increase of the leaf area index.Setting parameters with lower sensitivity to fixed values,filtering parameters sensitive to reflectance in the red and near-infrared bands to participate in the construction of a simulated data set,and stepwise setting of the step of the leaf area index can effectively improve calculations effectiveness.(3)A look-up table was established based on the analysis of the vegetation reflectance-LAI relationship to achieve inversion of leaf area index.From the perspective of the triangulation of the cap,the vegetation reflectance-LAI relationship is universal,and the relationship between the two vegetations is different.Full understanding of the reflectivity of all types of vegetation-the LAI relationship helps to construct a more efficient look-up table that can improve the accuracy of inversion.The accuracy of the look-up table in this study was evaluated using measured data.The RMSE of leaf area index look-up tables for maize,rice,desert scrub,and grassland were 0.47,0.43,and 0.51,0.53,respectively.The accuracy of the farmland leaf area index retrieval was slightly higher than that of natural growing desert vegetation.(4)The accuracy of using the WorldView-3 remote sensing image leaf area index inversion is high.WorldView-3 remote sensing image has high spatial resolution,and more pure pixels have played a positive role in the retrieval of leaf area index.The use of vegetation extraction results as ancillary data in the inversion process,sub-category inversion of leaf area index,can improve the accuracy.