Back to Projects
AgriGuard: AI-Powered Crop Disease Detection
AI-powered crop disease detection system built for Ag Leader after collecting requirements directly from their team.
Overview
Built for Ag Leader, an agriculture industry leader, after collecting requirements directly from the field. Every feature maps to a real farm problem.
Problem
Crop disease identification takes days when waiting for a field expert visit, leading to potential crop loss.
Solution
An AI-powered mobile application and IoT hardware integration that flags diseases in seconds and automatically manages irrigation and alerts.
Features & Architecture
- Machine Learning: Trained a MobileNetV2 transfer learning model across 39 crop disease classes with an EfficientNetB0 plant gate that rejects non-crop images before inference. Achieved 95% accuracy.
- Backend: Built a Node.js/Express REST API with JWT auth, scan history, and a pesticide database so the app tells the farmer what to spray and when.
- Hardware Integration: Wired an ESP32-CAM for automated image capture and a relay module for water pump control. Both are manageable remotely through the mobile app.
Results
Disease identification reduced from days to seconds. No on-site intervention needed for basic irrigation management.