| Leaf area index(LAI),an important descriptive parameter of forest structure canopy,has become a research hotspot in recent years.There was a problem of saturation point of vegetation index when extracting leaf area index by vegetation index.The vegetation index was easy to reach saturation in areas with high LAI,and the vegetation index no longer changes with increasing LAI,which would lead to the underestimation of LAI.This study analyzes the mixed image elements constituted by the vegetation-soil system based on Bell’s law.The normalized vegetation index(NDVI),ratio vegetation index(RVI),soil-adjusted vegetation index(SAVI),transformed soil-adjusted vegetation index(TSAVI),anti-atmospheric vegetation index(ARVI),modified soil-adjusted vegetation index(MSAVI),enhanced vegetation index(EVI),Perpendicular vegetation index(PVI),and differential vegetation index(DVI)were selected as the nine vegetation indices to construct the VI-LAI functional relationship.Simulate the canopy reflectance of broadleaf and conifer trees with four types of leaf inclination distribution:uniform distribution,spherical distribution,flat-like distribution,and straight-like distribution,and substitute the reflectance into the VI-LAI functional relationship.Analysis of saturation of VI with LAI for broadleaf and conifer trees with different leaf inclination angles.And Maoershan experimental Forest Farm as the research area,using field measurement data verification.The research results showed that RVI had the highest sensitivity to LAI and had a linear relationship with LAI after theoretical analysis of the selected 9 planting cover indices and LAI.ARVI,TASVI and NDVI showed better effects;EVI,SAVI and MSAVI had poor sensitivity to LAI.PVI and DVI had the lowest sensitivity to LAI.Based on the simulation data and the results of sensitivity analysis,RVI was linearly related to LAI at uniform distribution of leaf inclination,spherical distribution and hi-rectangular distribution,and the other 8 vegetation indexes gradually saturated with the increase of LAI,and the saturation points of different vegetation indices were different.In the planting quilt index,RVI showed the best performance,the highest sensitivity to LAI and the strongest saturation resistance.ARVI,NDVI and TSAVI were the next most sensitive to LAI.SAVI,EVI and MSAVI had general applicability,and had low sensitivity to LAI,and SAVI sensitivity was higher than MSAVI.PVI and DVI showed the worst performance,the lowest sensitivity to LAI and the weakest saturation resistance,which was the same as the theoretical analysis results.The red band reflectance of broadleaf and coniferous vegetation canopy gradually saturated with the increase of vegetation density,resulting in the vegetation index calculated by the band reflectance also tends to be saturated,and different vegetation indices have different saturation points due to different calculation formulas.Through fitting and verification of vegetation index and field measured LAI data,RVI and measured LAI showed a linear relationship,which was the same as the results of theoretical analysis and simulation data.The other 8 vegetation indexes were saturated with the increase of measured LAI,which was the same as the simulation data.ARVI was highly correlated with measured LAI,followed by TSAVI and NDVI.ARVI gradually saturated after LAI approached 6,while NDVI and TSAVI were in LAI approached 5 gradually saturated,EVI,SAVI and MSAVI have low sensitivity to measured LAI.SAVI sensitivity was higher than MSAVI sensitivity.PVI and DVI had the lowest sensitivity.The ranking of vegetation index sensitivity from high to low was the same as the simulation data.Therefore,RVI was selected as the best suitable vegetation index for LAI inversion in the study area,and compared with ARVI,RVI-LAI linear model and ARVI-LAI index model were established.It has been verified that RVI-LAI model inversion results have the highest accuracy,R~2=0.844.So RVI can be used for regional LAI inversion quickly and accurately. |
PDF Full Text Request