High Resolution Multi-sensor Data Comparison for Mapping Mango Orchards

Mango, the king of fruits is an important part of the Indian economy and household. For obtaining a good yield it is important to apply sustainable and efficient orchard management practices. This study compares two high resolution datasets; LISS-4 and Sentinel-2. LISS-4 dataset has high spatial resolution and Sentinel-2 has high spectral resolution. Mapping was carried out using object-based image analysis for classification. Three vegetation indices were used for classification; NDVI (Normalised Difference Vegetation Index), LSWI (Land Surface Water Index) and Red Edge Normalised Difference Vegetation Index (ReNDVI). NDVI was obtained from LISS-4 and Sentinel-2 imagery. The overall accuracy of level-2 classification using NDVI based LISS-4 imagery gave an overall accuracy of 87.2% with a kappa coefficient of 0.83. Level 3 classification of the same image gave an overall accuracy of 79% with a kappa coefficient of 0.74. Out of all the indices used for level 2 classification with Sentinel-2, ReNDVI performed the best with an overall accuracy of 81.6% and a kappa coefficient of 0.76. In the case of level-3 classification, LSWI gave the best performance with an overall accuracy of 80% and a kappa coefficient of 0.77.