Mapping Orchards and Crops Using Sentinel-2 Imagery

Abstract

The mapping and identification of crops and orchards in area is important for forecasting crop yield, evaluating the factors inducing the crop stress, management of resources as well as for the formulation of policy. In Indian Punjab, crop diversification is focused for sustainable agriculture and promoting less water intensive crops. The cultivation of horticultural crops has emerged as one of the viable alternatives for diversification from current paddy-wheat cropping system. Therefore, different orchards were mapped in the Abohar Tehsil of Punjab using Sentinel-2 satellite data (August 2021) which was classified using iso-cluster unsupervised classification technique. Paddy and cotton were differentiated in the near-infrared band (NIR) and orchards from other features using green band. Among different orchards in the area, Kinnow (citrus fruits) was differentiated only from other orchards (like guava, orange, malta and ber) and is cultivated in 13.2% area during 2021. Overall accuracy for classification was 94.1% with the kappa coefficient of 0.87. These results showed that Sentinel-2 can be used for mapping of orchards and monitoring crop diversification.