🔍 PiVision: Object and Animal Recognition with Raspberry Pi

📝 Description

PiVision is a sophisticated computer vision project that leverages the power of Raspberry Pi to identify and track objects and animals in real-time. Using the COCO dataset, this system can recognize 91 unique objects/animals while providing constantly updating confidence ratings for each detection.

✨ Features

• Real-time object and animal recognition
• Detection of 91 different object classes from the COCO dataset
• Live confidence rating updates
• Optimized for Raspberry Pi hardware
• Lightweight implementation for edge computing

🛠️ Technologies Used

📸 Project Gallery

🚀 Installation

# Clona el repositorio
git clone https://github.com/JonIbarreche/PiVision.git

# Navega al directorio del proyecto
cd PiVision

# Instala virtualenv si no lo tienes
pip3 install virtualenv

# Crea un entorno virtual
python3 -m virtualenv env

# Activa el entorno virtual
source env/bin/activate

# Ahora instala las dependencias en el entorno virtual
pip install -r requirements.txt

# Configura la cámara de Raspberry Pi
sudo raspi-config
# Navega a "Interface Options" > "Camera" y habilita la cámara

# Asegúrate de que tu Raspberry Pi está actualizada
sudo apt-get update
sudo apt-get upgrade

💻 Usage

# Run the main recognition script
python main.py

🔧 Requirements

• Raspberry Pi (3 or newer recommended)
• Pi Camera or USB webcam
• Python 3.6+
• OpenCV
• TensorFlow Lite

👥 Contributors

• Jon Ibarreche jon.ibarreche@opendeusto.es
• Javier Fuente javifuenn@opendeusto.es