Modelling of surface roughness on agriculture area using Radarsat-2 satellite

Abstract

This research aims to model surface roughness for agricultural land using RADARSAT-2 satellite imagery. To obtain the model of surface roughness, the digital number of radar image are converted into backscattering coefficient and used as basis of surface roughness modeling. Furthermore, using the relationship between the backscattering coefficient, local incidence angle, and the wavelength was made initial models for surface roughness. The new equation that can model surface roughness on agricultural land is obtained by performing calibration between the value of the surface roughness from initial models and combinations of initials models with the value of surface roughness from field measurement that the equation resulted y=-591987x5+(9.106)x4-(6.107)x3+(2.108)x2-(3.108)x+(2.108) with RMSE value = 0.31 cm, where y is the value of surface roughness modeled (cm) and x is the surface roughness value of the combination of initial models used during the calibration process (cm). The combination of HV, VH polaristation and local incident angle (h0HV+h0VH)x cos ϴ produces a correlation value (R2) = 0.804.