Research On Vegetation LAI Inversion Under Correction Of Needle Leaf Inclination

With the increasing degree of globalization in our country,the high-grade highways continue to promote rapid economic development,but the construction and operation of highways will inevitably have a greater impact on the road environment,such as reducing vegetation coverage,soil erosion,etc.,so it’s very necessary to monitor the highway environment quickly and accurately.In this study,the leaf area index(LAI),which represents the growth status of vegetation,was taken as the research objective to provide a basis for monitoring the dynamic changes of road environment.Aiming at the problem that the leaf inclination angle distribution type in the canopy SAIL model is difficult to apply to the needle leaf inclination angle distribution in this study,the needle leaf inclination angle is corrected on the basis of the existing SAIL model,and the ESAIL model is obtained.Then the study combins with the needle leaf radiation transmissin model LIBERTY to build a LIBERTY+ESAIL coupling model for LAI inversion.The ChangYi(Changsha-Yiyang)Expressway in Hunan Province is used as the research road area,GF-6 remote sensing images,model simulation data and field-measured parameters during the same period are used as data sources.The LIBERTY+EESAIL model is used for multiple linear regression and partial the weighted regression is used to quantitatively invert LAI,and the inversion results are compared with the results,whitch are LIBERTY+SAIL model and the LAI inversion results of the two regression methods.The foremost research conclusions of my study are as follows:(1)The canopy SAIL model needle leaf inclination correction.On the basis of fully studying the characteristics of needles and the distribution types of leaf inclination angles,the paper combines with the distribution of needle leaf inclination angles,the parameter-adjusted two-parameter ellipsoid distribution function is embedded in the SAIL model to obtain the ESAIL model and realize the needle leaf inclination angle of SAIL model correction.At the same time,the accuracy of the ESAIL model is evaluated.Then the accuracy of the spectrum simulation is improved after model revised.(2)Determine the inversion factors of the inversion model.In this study,the single-band reflectances of GF-6 images are extracted,and the correlation among the ratio vegetation index RVI,the soil regulation index SAVI,the environmental vegetation index EVI,other five vegetation indices and the four single-band reflectances with LAI are analyzed.The results show that the modeling accuracy of the eight-planted quilt indices is generally higher than the modeling accuracy of the four single-band reflectances.According to the correlation results,the three vegetation indices of SAVI,RVI and EVI with higher correlation are used as the inversion factors of the coupled model.(3)Construct the combined model of the coupling model and the two regression methods to invert LAI.In this study,a combination model of LIBERTY+SAIL,multiple linear regression and local weighted regression are constructed to invert LAI.Subsequently,a combination model of LIBERTY+ESAIL,multiple linear regression and local weighted regression are constructed to invert LAI.The comparative analysis of the four inversion results shows that the LIBERTY+ESAIL model has a better prediction effect on LAI,and the prediction effect of the combined model and machine learning combination is better than the same multiple linear regression combination.It can be seen that the combination of the LIBERTY+ESAIL coupling model and the local weighted regression method to invert LAI has certain advantages.The results show that the combined ESAIL model and LIBERTY model after correction of the conifer inclination angle distribution function and local weighted regression has good predictive ability for LAI,which can be the dynamic monitoring of road vegetation health in typical hilly areas in southern China and provide basic support.