Project Overview
Agron is an AI-driven autonomous agricultural drone system designed to revolutionize precision farming through intelligent crop monitoring and targeted intervention. The system integrates multispectral imaging, advanced computer vision, and autonomous flight control to enable early detection of crop stress and disease, followed by precision pesticide application only to affected areas.
Traditional agricultural practices suffer from delayed pest detection, excessive chemical usage, and inefficient field monitoring. Agron addresses these challenges by combining RGB and Near-Infrared (NIR) cameras for NDVI-based vegetation analysis, a mobile Ground Control Station for intuitive mission planning, and Raspberry Pi edge computing for offline operation in rural environments. This integrated approach reduces chemical usage by over 30% while improving crop health monitoring and operational efficiency.
Technical Architecture
Hardware Components
- Pixhawk Autopilot (ArduPilot Firmware)
- Raspberry Pi 5 (Companion Computer)
- RGB Camera (Visual Inspection)
- NIR Camera (Vegetation Analysis)
- Multirotor UAV Platform
- Autonomous Spraying Unit
- MAVLink Serial Communication
- Android Mobile GCS Device
Software & AI
- Flutter Mobile GCS (Dart)
- Python Backend (FastAPI/Flask)
- pymavlink for MAVLink Protocol
- Firebase (Auth & Cloud Sync)
- NDVI Vegetation Index Analysis
- CNN & YOLO-tiny for Detection
- Google Gemini API (AI Assistance)
- WebSocket Real-time Communication
AI & Computer Vision Pipeline
Multispectral Imaging & NDVI Analysis
Agron employs synchronized RGB and Near-Infrared (NIR) cameras to capture multispectral imagery during autonomous flight. The system computes the Normalized Difference Vegetation Index (NDVI) to quantify crop health, enabling early identification of stressed areas before visible symptoms appear. This allows for proactive intervention and targeted treatment.
Computer Vision Techniques
- • Feature-based image alignment
- • Image stitching and mosaicking
- • NDVI computation from RGB/NIR
- • Threshold-based stress detection
- • K-Means/DBSCAN region clustering
AI/ML Models
- • CNN classifiers (MobileNet)
- • YOLO-tiny for object detection
- • Pest and disease identification
- • Stress zone classification
- • Optimal path calculation
Key Features
Autonomous Flight Control
Pixhawk-based autonomous navigation with MAVLink protocol for reliable waypoint execution and mission management.
Targeted Precision Spraying
AI-driven spraying system that applies pesticides only to detected stressed zones, reducing chemical waste by 30%+.
Multispectral Imaging
Synchronized RGB and NIR cameras for comprehensive crop health assessment through NDVI vegetation index.
Mobile Ground Control Station
Flutter-based cross-platform GCS with intuitive map-based mission planning and real-time telemetry visualization.
AI-Assisted Mission Planning
Google Gemini API integration for natural language mission input and intelligent flight path generation.
Offline-First Architecture
Raspberry Pi edge computing enables full system operation without internet connectivity for rural deployment.
Impact & Results
Project Achievements
Agron successfully demonstrates an integrated precision agriculture workflow combining autonomous flight, multispectral crop analysis, and targeted intervention. The system validates the feasibility of synchronized image acquisition, edge-based NDVI computation, and AI-assisted mission planning for sustainable farming operations.
- 30%+ reduction in chemical usage
- Early pest/disease detection via NDVI
- Autonomous mission execution validated
- Offline-first rural deployment
Technical Achievements
- ✓ Multispectral NDVI analysis
- ✓ Pixhawk MAVLink integration
- ✓ Flutter cross-platform GCS
- ✓ Python FastAPI backend
- ✓ Edge-based image processing
System Capabilities
- → Targeted precision spraying
- → Real-time telemetry monitoring
- → AI-assisted mission planning
- → Offline operation capability
- → Crop stress visualization
Sustainability & Environmental Impact
Chemical Reduction
Targeted spraying reduces pesticide usage by 30%+, minimizing environmental pollution and groundwater contamination.
Soil Health
Precision application protects beneficial soil microbiota and maintains long-term soil fertility.
Sustainable Farming
Supports global sustainability goals including clean water, responsible consumption, and climate action.