Estimation of Wheat Crop Production using Multispectral Information Fusion.

Abstract

The present work estimates the area and the corresponding produce of wheat crop in the study area, which comprises Etah region of Uttar Pradesh, India. For this purpose, multispectral images of multiple sensors, viz., Sentinel-2, Landsat-8, and Landsat-9 during the preharvesting period, i.e., March for the years 2021 and 2022 are used. A multispectral information fusion approach has been proposed which involves image classification as well as vegetation index-based information extraction. For imposing the information fusion, appropriate image bands are identified with the help of separability analysis followed by land cover classification for wheat crop class extraction. Support Vector Machine (SVM), Artificial Neural Network (ANN) and Maximum Likelihood (ML) are used for classification whereas Normalized Difference Vegetation Index (NDVI) and Fractional Vegetation Cover (FVC) are used for index-based crop area extraction. The maximum accuracy of 98.34% is achieved for Sentinel-2 data using ANN whereas minimum accuracy of 80.21% is achieved for Landsat-9 using ML classifier. The estimated area for Sentinel-2 data for the year 2021 is 260540 hectares using ANN and 203240 hectares using ML which is near to the reference data, i.e., 238600 hectares. SVM also shows good performance and calculates least error in estimated crop area for the year 2022 on Sentinel-2 data. It calculates 8408490 tons of wheat for the same year. The proposed method utilizes a single image per year for extraction of information supported by the ground truth data which makes it a novel approach of information extraction for crop produce monitoring. This article is protected by copyright. All rights reserved.