A real-time rice quality detection system using ESP32-CAM and YOLOv8 object detection. The system captures images of rice grains, sends them to a Flask server for AI-powered quality analysis, and activates a relay to automatically sort defective grains.

• Real-time Detection: Continuous monitoring and classification of rice grains
• AI-Powered Analysis: YOLOv8 model trained to detect 5 types of rice defects
• Automated Sorting: Relay-controlled mechanism to separate bad rice
• Web Monitoring: WebSerial interface for remote monitoring and debugging
• Low Latency: Optimized for fast inference and response times

The system can identify the following rice quality issues:

1. Damaged - Physically damaged grains
2. Discolored - Color abnormalities
3. Broken - Fractured or incomplete grains
4. Chalky - Opaque or chalky appearance
5. Organic Foreign Matters - Non-rice contaminants

• ESP32-CAM (AI-Thinker module)
• Relay Module (for sorting mechanism)
• Power Supply (5V for ESP32-CAM)
• USB-to-Serial Adapter (for programming)
• Optional: Flash LED (built-in on GPIO 4)

Project-ESPCAM/
├── esp32_cam_firmware/          # ESP32-CAM Arduino code
│   ├── rice_detection_esp32cam.ino
│   ├── config.h.template        # Configuration template
│   └── config.h                 # Your actual config (gitignored)
├── server/                      # Flask API server
│   ├── app.py
│   └── requirements.txt
├── models/                      # Trained YOLOv8 models
│   └── best.pt
├── dataset/                     # Training dataset
│   └── rice_object_detection.v1i.yolov8/
├── docs/                        # Documentation
│   ├── HARDWARE_SETUP.md
│   ├── DEPLOYMENT.md
│   └── API_DOCUMENTATION.md
└── README.md

# Navigate to server directory
cd server

# Install dependencies
pip install -r requirements.txt

# Run the Flask server
python app.py

The server will start on http://0.0.0.0:5000

1. Configure WiFi and Server:

   cd esp32_cam_firmware
   cp config.h.template config.h

2. Edit config.h with your settings:

   const char *WIFI_SSID = "YourWiFiName";
   const char *WIFI_PASSWORD = "YourPassword";
   const char* SERVER_URL = "http://YOUR_SERVER_IP:5000/predict";

3. Upload to ESP32-CAM:

   • Open rice_detection_esp32cam.ino in Arduino IDE
   • Select Board: "AI Thinker ESP32-CAM"
   • Select Port: Your USB-to-Serial adapter port
   • Click Upload

4. Monitor via WebSerial:

   • After upload, connect ESP32-CAM to power
   • Open Serial Monitor to see the IP address
   • Navigate to http://ESP32_IP/webserial for web-based monitoring

• Hardware Setup Guide - Detailed wiring and assembly instructions
• Deployment Guide - Server deployment options (local, cloud)
• API Documentation - Flask API endpoints and usage

Edit esp32_cam_firmware/config.h:

Edit server/app.py:

1. Image Capture: ESP32-CAM captures an image of rice grains
2. Transmission: Image is sent via HTTP POST to the Flask server
3. AI Analysis: YOLOv8 model processes the image and detects defects
4. Response: Server returns JSON with detection results and bad_rice_detected flag
5. Sorting: If bad rice is detected, ESP32-CAM activates the relay to trigger sorting mechanism
6. Repeat: Process continues every 10 seconds

• Image Capture: ~50-150ms
• Network Transmission: Varies by connection
• Model Inference: ~100-500ms (depends on server hardware)
• Total Cycle: ~10 seconds (configurable)

Contributions are welcome! Please feel free to submit issues or pull requests.

This project is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License - see the LICENSE file for details.

Copyright (c) 2026 Atanda Isaac-Great Peace and Ilori Joshua Ayomide

For commercial use or derivative works, please contact: [email protected]

• Dataset provided by Roboflow Universe
• YOLOv8 by Ultralytics
• ESP32-CAM community and resources

For questions or issues:

• Open an issue on GitHub
• Check the documentation
• Review existing issues and discussions

Made with ❤️ for agricultural quality control