Advanced mapping of mango orchards in Chittoor district through object-based image analysis approach using Sentinel-2 imagery

Mango is crucial in India's agricultural economy, particularly in the Chittoor District of Andhra Pradesh. As mango cultivation transitions from subsistence farming to large-scale commercial operations, accurate mapping and monitoring of mango plantations are essential for sustainable agricultural management. This study explored the application of Object-Based Image Analysis (OBIA) using high-resolution Sentinel-2 satellite imagery for mapping mango plantations in Chittoor District. OBIA, a more advanced approach than traditional pixel- based methods, integrates spectral, spatial and contextual data, enabling the identification of mango orchards with higher accuracy. Sentinel-2 multispectral bands were utilized to distinguish mango plantations based on canopy density, inter-row spacing and orchard layout. Ground-truth data collected from 531 points across the district validated the classification process. The results show that OBIA achieved an overall accuracy of 87.0 % with a Kappa index of 0.74, signifying strong agreement with the ground truth data and the total mango area mapped in Chittoor District is 97,006 hectares. This study highlights the effectiveness of OBIA for accurate mango area estimation and suggests potential improvements, including the integration of hyperspectral data and advanced algorithms. This study offers valuable insights for agricultural management, resource optimization and policy planning, with implications for broader crop mapping and precision agriculture applications.