Semi-empirical SAR vegetation index for crop discrimination based on biomass in semi-arid region: A case study in Perambalur district, India

Abstract

Crop discrimination helps monitor crops and strengthen the agricultural industry to secure food supplies. From existing crop-related studies and observations, found that vegetative Empirical models were limited in crop classification for Kharif and Rabi season crops based on biomass. To improve biomass assessment, this study proposes a modified Vegetation Index using the deviation factor from the plot between the Dual Polarization SAR Vegetation Index (DPSVI) and the Dual Polarimetric Radar Vegetation Index (DpRVI). The deviation factor corrects the difficulties in the lower and higher ranges and is found to be 0.5 for Kharif and Rabi at the harvest stage. Performance of Integrated SAR Vegetation Index carried out to both seasons and shows Kharif (R2 = 0.83), Rabi (R2 = 0.87). In addition, cross-validation with Sentinel-2 derived Normalized Difference Vegetation Index, which has R2 values of 0.84 and 0.86 for Kharif and Rabi seasonal crops respectively. A correlation was performed to assess the capability of the proposed model for individual crops and found that good correlation with vertically grown Kharif and rabi crops except for Sesamum Indicum. A classification capability of the Kharif (76%) and Rabi (84%) crops and variation shows crops are healthy and have more greenery. From the results, a modified measure outperformed than DPSVI and DpRVI at the harvest stage for a selected crop in the above season and proved an alternate.