A hackathon prototype for real-time waste detection and sorting using computer vision, Arduino, and a web dashboard.

Smart Bin is an end-to-end system that classifies waste in real-time, actuates hardware to sort it, and provides a web-based interface for monitoring and control. It demonstrates edge-to-cloud integration with AI models, embedded hardware, and a responsive UI.

• Real-time waste classification: Detects categories like recyclable, compost, and landfill.
• Lightweight CNN model: Trained with TensorFlow/Keras; GPU-compatible.
• Edge inference: Runs on Raspberry Pi or other lightweight devices.
• Arduino integration: Controls servo/motor for automated bin sorting.
• Web dashboard: Monitor camera feed, see classification results, and view system logs.
• Data pipeline: Utilities for dataset preparation and augmentation.
• Reproducible deployment: Dockerfile/scripts for easy setup.

• Backend/Model Training: Python, TensorFlow/Keras, OpenCV
• Hardware: Arduino, Raspberry Pi (camera input)
• Frontend: HTML/CSS/JavaScript (Tailwind optional), React (optional)
• DevOps: Docker

• https://github.com/belanasaikiran/Smart-E-Waste-Bin-Sorter

# create a virtual environment
python -m venv .venv
source .venv/bin/activate  # Linux/Mac
.venv\Scripts\activate     # Windows

# install dependencies
pip install -r requirements.txt

# run inference server
python src/inference_server.py --model models/smartbin_v1.h5 --camera 0

# access web UI
# Open http://localhost:3000 in your browser
# Smart Bin — HackUMass 2025

A hackathon prototype for real-time waste detection and sorting using computer vision, Arduino, and a web dashboard.

Smart Bin is an end-to-end system that classifies waste in real-time, actuates hardware to sort it, and provides a web-based interface for monitoring and control. It demonstrates edge-to-cloud integration with AI models, embedded hardware, and a responsive UI.

• Real-time waste classification: Detects categories like recyclable, compost, and landfill.
• Lightweight CNN model: Trained with TensorFlow/Keras; GPU-compatible.
• Edge inference: Runs on Raspberry Pi or other lightweight devices.
• Arduino integration: Controls servo/motor for automated bin sorting.
• Web dashboard: Monitor camera feed, see classification results, and view system logs.
• Data pipeline: Utilities for dataset preparation and augmentation.
• Reproducible deployment: Dockerfile/scripts for easy setup.

• Backend/Model Training: Python, TensorFlow/Keras, OpenCV
• Hardware: Arduino, Raspberry Pi (camera input)
• Frontend: HTML/CSS/JavaScript (Tailwind optional), React (optional)
• DevOps: Docker