Regional scale inversion of chlorophyll content of Dendrocalamus giganteus by multi-source remote sensing

[Objective]The spectrophotometer method is costly,time-consuming,laborious,and destructive to the plant.It will be lost during the transportation process,and can only obtain sample point data,which is a challenge to the estimation of chlorophyll content at the regional level.[Method]In this study,in order to improve the estimation accuracy,a new method of collaborative inversion of chlorophyll using Landsat 8 and Global Ecosystem Dynamics Investigation(GEDI)is proposed.Specifically,the chlorophyll content data set is combined with the preprocessed two remote sensing(RS)factors to construct three regression models of support vector machine(SVM),BP neural network(BP)and random forest(RF),and the better model is selected for inversion.In addition,the Kriging method is used to interpolate the GEDI point attribute data into the surface attribute data for modeling.[Result]The results showed that:(1)The chlorophyll model of single plant was y=0.1373x1.7654.(2)The optimal semi-variance function models of pai,pgap_theta and pgap_theta_a3 are exponential models.(3)The top three correlations between the two RS data and chlorophyll content were B2_3_SM,B2_3_HO,B2_5_EN and pai,pgap_theta,pgap_theta_a3.(4)The combination of Landsat 8 image and GEDI had the highest modeling accuracy,and RF had the best performance,with R2,RMSE and P of 0.94,0.18 g/m2 and 83.32%,respectively.[Conclusion]This study shows that it is reliable to use Landsat 8 images and GEDI to retrieve the chlorophyll content of Dendrocalamus giganteus,and reveals the potential of multi-source RS data in the inversion of forest ecological parameters.