Inversion Of Leaf Area Index Of Rice Based On Radiative Transfer Model And Landsat ETM+Data

Leaf Area Index(LAI),one-sided leaf area unit of ground area,is an importantstructure parameter of vegetation ecosystems, which have relationship with conditionof the vegetation, and it’s also a key parameter of many vegetation-atmosphereinteraction model, in particular on the model of carbon and water cycles. Since mostLAI products didn’t take into account the rice’s growing environment,and set the soilbackground too simple, so this will inevitably lead to the error.To solve these problems, we use Landsat ETM+remote sensing data of11/5/2003in the study area, combined with radiative transfer model (PROSAIL),toinverse LAI of rice. Finally, we verified the inversion results according to the in-situdata of10/25/2003. In this paper, research work include:1) Data collection and pre-processing. It include the downloading and theatmosphere correction of the ETM+remote sensing data,collection andcollation of the in-situ data.2) Based on the results of the parameters’ sensibility analysis of the PROSAILparameter, besides the changes of the rice in this period, four sensitiveparameters were chosen,which include Chlorophy II a+b content, Equivalentwater thickness, LAI and soil reflectance. And validate the forwardsimulation was validate by compare the simulation reflectance and the filedreflectance.3) Based on the PROSAIL model and by change the sensitive parameter withcertain step,we simulation18000reflectance data set to prepare for thetraining data and the validation data in the artificial neural network.4) The inversion model was built by training the simulation data usingBack-Propagation Network (BP network).And then the model is used toinverse the ETM+LAI.5) Validate the inversion results by the in-situ data. And the results showed thatthe model built by artificial neural network(ANN) can reflect the distributionof the LAI, however, the inversion LAI is smaller than the field LAI.