Inspiration
Agriculture is a backbone of India’s economy, yet farmers face uncertainty in crop planning due to fragmented farms, variable climate, and limited access to timely information. Manual surveys are slow and expensive, often leading to inefficient resource allocation. We were inspired to explore how satellite data and AI could provide accurate crop type identification, helping farmers, policymakers, and agri-businesses make informed decisions.
What it does
SmartCrop aims to: Classify crop types (e.g., rice, wheat, maize, cotton) using multi-temporal satellite imagery. Use SAR and optical satellite data along with vegetation indices (NDVI, EVI). Perform time-series analysis to differentiate crops across growth stages. Provide a dashboard: users can input coordinates to see farm details such as crop type, growth stage, and stress indicators. Optional: flag potential crop stress or anomalies for early action.
How we built it
(Planned approach, since implementation hasn’t started) Data sources: Sentinel-2, Landsat-8, AgriFieldNet India. Preprocessing: Cloud removal, temporal alignment, vegetation index computation. Models: Random Forest, XGBoost, CNNs, or transformer-based time-series models. Dashboard: Flask backend, React.js frontend, Folium for interactive maps
Log in or sign up for Devpost to join the conversation.