Paper Key : IRJ************784
Author: Chaudhari Tanmay Prabhakar,Umesh Gangadhar Sake,Vinanti Ghanshyam Shinde,Srushti Vijay Ghise,Nikita Umesh Girase
Date Published: 13 Nov 2024
Abstract
Precision agriculture has become a vital area in recent years for tackling the problems of sustainability and global food security. Yield prediction is a crucial component of precision agriculture that aids farmers and other agricultural stakeholders in making well-informed choices on crop management, resource allocation, and strategic planning. To increase crop health and productivity, precise yield prediction enables timely interventions, focused fertilizer application, and optimal irrigation scheduling. Conventional yield estimation techniques rely on historical data and oversimplified statistical models, which frequently fall short in capturing the intricacies of crop growth influenced by management techniques, soil conditions, and environmental factors.Advancements in machine learning have provided new avenues for improving yield prediction accuracy. Remote sensing data, particularly multispectral imagery, captures critical spatial and spectral information about crops, offering insights into plant health, growth stages, and environmental stress factors. Machine learning models, especially deep learning architectures, have shown great promise in leveraging this data to reveal complex patterns that traditional approaches may overlook.However, existing deep learning models typically address spatial or temporal dimensions independently, which limits their ability to capture the multifaceted dynamics of crop growth over time. Experimental results demonstrate that the proposed model outperforms traditional yield prediction methods, achieving lower root mean square error and higher correlation coefficients.
DOI Requested