Leaf Area Index Inversion From Multi-angle Remote Sensing

Leaf Area Index(LAI)reflects the structural characteristics and biophysical changes of vegetation canopy in the ecosystem,and has been widely used in the study of ecosystem carbon cycle,global climate models and many other terrestrial ecological biological models.Therefore,how to obtain large-area and high-quality LAI quickly and in real time has become one of the hot issues in quantitative remote sensing research.At present,LAI inversion using remote sensing data is limited by the limitations of inversion methods and insufficient observation data,resulting in low accuracy of inversion results.To solve this problem,we can make full use of the rich ground object structure information provided by multi-angle remote sensing observation data,and improve the inversion method from the traditional single zenith direction to multi-angle interpretation information.In this study,a LAI inversion method suitable for multi-angle remote sensing was proposed.Based on PROSAIL model and nuclear driven BRDF model widely used in vegetation inversion,multi-angle LAI was retrieved using LANDSAT multi-spectral reflectance data and MODIS BRDF product(MCD43A1).By comparing MODIS LAI product(MOD15A2H)and ground measured LAI data,the feasibility of inversion method and accuracy of inversion results were verified.The main research contents and conclusions are as follows:(1)The sensitivity of LAI was analyzed based on PROSAIL model.Two methods,RMSE and EFAST,were used to quantitatively evaluate the sensitivity of LAI to model simulated reflectance on parameters and angles,and it was confirmed that LAI changes would cause significant changes in simulated reflectance,which were mainly manifested in hot spots and their nearby geometric locations.At the same time,considering the BRDF effect of PROSAIL model,the influence mechanism of LAI on the three kernel coefficients of the nuclear drive model was further analyzed,and the results were all highly sensitive.(2)Coupling PROSAIL model and nuclear drive model,and improving their coupling differences.By comparing the differences between the two models in the simulated multi-angle reflectance,it is found that the model coupling has a good consistency,but the difference is great near the hot spot.In order to improve the coupling difference between the models,the hot spot effect of the core driven model is discussed.Based on existing studies,we proposed a new idea to delineate hot spot factors with different solar zenith angles,and finally introduced the improved RTLSR?C model of hot spot factors,which significantly improved the simulated reflectivity of hot spot direction and effectively optimized the coupling effect of the model.(3)Inversion of multi-angle LAI based on PROSAIL model and nuclear drive model.The relationship lookup table between sufficient LAI sampling values and multi-angle reflectance simulated by PROSAIL was established,and the statistical relationship between NDVI and LAI was established with the help of the standby algorithm of MODIS LAI products.The possible LAI value range was preliminarily screened by comparing and calculating surface OBSERVED NDVI,which effectively improved the inversion accuracy and speed.LAI in the zenith direction was retrieved by using LANDSAT reflectance data,and multi-angle reflectance was fitted by using MODIS BRDF products through model coupling,so as to invert LAI in the zenith,cold and hot spots directions.Finally,the accuracy was verified by comparing MODIS LAI products with ground measured data.The results show that LAI inversion results have good spatial continuity,and the inversion accuracy of cold and hot spots is higher than that of single zenith direction.Therefore,the method of using multi-angle remote sensing to invert LAI in this paper is feasible and can improve certain inversion accuracy.This method can be further applied to LAI inversion at any Angle.