 |
Online ISSN : 2349-8080 Issues : 12 per year Publisher : Excellent Publishers Email : editorinchiefijcrbp@gmail.com |
Accurate crop area estimation is essential for agricultural planning and food security. This study explores the integration of Sentinel-1A SAR and Sentinel-2A optical data to map kharif paddy areas in Tamil Nadu’s Cauvery Delta region. Data acquired from May to October 2019 were analyzed using parameterized classification of SAR backscatter time-series and maximum likelihood classification of optical imagery. Ground truth data from 300 field samples supported model training and validation. While SAR provided reliable temporal data unaffected by cloud cover, optical data offered rich spectral information. Pixel-level agreement between the two classification outputs was assessed using TerrSet's ERRMAT module, yielding a Kappa value of 0.83. Feature-level fusion was then applied, resulting in an integrated paddy map estimating 86,152.20 hectares across Thanjavur, Thiruvarur, and Nagapattinam districts. Accuracy assessment using independent validation points revealed an overall accuracy of 94%, with rice classified at 96% accuracy and a Kappa index of 0.86. The integration approach improved accuracy over individual datasets. This study confirms that combining SAR and optical data enhances classification robustness and supports operational-scale crop monitoring under frequent cloud cover, providing valuable input for precision agriculture and policy planning.
