Application of Geospatial Technology for Wheat Crop Identification and Crop Coverage in Akola District of Maharashtra, India

Abstract

Accurate and timely crop-area information is required for agricultural planning, irrigation management, and production forecasting. This study evaluated multi-date Sentinel-2A imagery and Normalized Difference Vegetation Index (NDVI) time-series analysis for identifying wheat and estimating its cultivated area in Akola district, Maharashtra. Sentinel-2A images acquired from November 2023 to April 2024 were processed using ERDAS Imagine and ArcGIS. Ground-truth observations were collected at 40 field sites between 30 December 2023 and 11 January 2024. NDVI values from wheat, gram, and other crops were used to develop Reference Temporal Spectral Profiles (RTSPs). An 11-date NDVI stack was classified through two-stage unsupervised ISODATA clustering, generating 50 spectral classes. Cluster-wise Temporal Spectral Profiles were compared with the RTSPs, and residual mixed pixels were interpreted visually to prepare the final classified image. Pixel enumeration at a spatial resolution of 10 × 10 m produced an estimated wheat area of 25,793 ha. Compared with the Department of Agriculture figure of 32,570 ha, the estimate was 20.81% lower. The findings indicate that temporal NDVI profiling can support district-scale wheat identification; however, the observed difference from official statistics and the absence of an independent error matrix indicate that further validation is required before operational application. Independent uncertainty assessment is therefore necessary.