Surface Soil Moisture Estimation Using a Neural Network Model in Bare Land and Vegetated Areas

Most of the approaches to retrieve surface soil moisture (SSM) by optical and thermal infrared (TIR) spectroscopies are purposed calculate various characteristic bands/indices then establish regression relationship between them in combination with measurement data. However, due combined impact many factors, often shows nonlinearity. Moreover, single temporal image measured data not transplantable time space, which makes it difficult construct a more general model for remote sensing (RS) estimation SSM. In order solve this problem, back propagation (BP) neural network (NN) an excellent nonlinear mapping ability is introduced determine band/index BPNN model, TIR RS different periods were taken as input parameters, situ treated output parameter. There 12 schemes designed The key findings study follows: (1) could SSM high accuracy that indicates correlation coefficient estimated 0.9001 (2) retrieval based on can be applied estimate spatial resolution values.