A secure remote monitoring system for fires, especially wildfires utilising TensorFlow, MongoDB and Auth0. It enables old devices to be turned into fire recognition sensors for an automated system which can be easily accessed using the simple web portal.

The Dataset

• Choose a dataset, such as this

• Make sure the directory of the images is separated into two subdirectories by class (Fire and Non Fire)

Training the model

• Set the directory of the images in the model.py file

• Run the model.py file

Run the main website

• Set your environment variables using .env.example as a base

• Run the website.py file

streamlit run "./streamlit/system.py"

streamlit run "./streamlit/aitester.py"