Paper 13083-63
Paper 13083-63
Harnessing satellite remote sensing for wheat lodging assessment in India: a machine learning approach
24 April 2024 • 13:30 - 15:00 Japan Standard Time | Exhibition Hall A
Abstract
This study investigates machine learning (ML) for wheat lodging detection in Haridwar, Uttarakhand (India), using Sentinel1 and Sentinel-2 satellite data from November 2022 to April 2023. Amidst limited research on ML applications in Indian agriculture, we focused on evaluating Random Forest (RF), Decision Tree (DT), and Support Vector Machines (SVM) algorithms based on ground control points and the Normalized Difference Vegetation Index (NDVI). RF emerged as the most accurate in classifying wheat areas, with significant wheat lodging observed in approximately 51% of the district's wheat fields. This approach, particularly the use of the Spectral Sum Index (SSI) for distinguishing lodged from non-lodged wheat, demonstrates ML's potential in enhancing crop yield assessments and informing decision-support systems.
Presenter
Indian Institute of Technology Roorkee (India)
Vishal Mishra is pursuing PhD at the Indian Institute of Technology Roorkee, India. He has to date published 9 journal papers, 9 conference papers 3 book chapters and one book. His research area includes the application of Remote Sensing for geohazards, water resources and urban mapping. He his working on the application of Artificial Intelligence on Geospatial data. He has been the recipient of the GRSG Student Award 2019 from the Geological Remote Sensing Group UK. He is also a recipient of the Helmholtz Visiting Researcher Fellowship from the Helmholtz Information & Data Science Academy (HIDA). He has worked as a visiting scientist at GFZ German Research Centre for Geosciences, Potsdam Germany for a few months.