Forest fires are one of the major reasons for the loss of forest ecosystems and human communities. More than 3 million forest land is destroyed every year. In the longer term, they can adversely affect the supply of environmental services, threaten the survival of endangered species, the composition and structure of forests, and the soil quality. Forest fires can be caused due to natural or human activities but whatever might be the cause they need to be prevented and managed properly to minimise the damage to the forest and animal ecosystem.

What it does

Fauna is an end-to-end solution to the above mentioned problems. We have three fold motto i.e prediction, prevention and restoration.

First we identify the forests susceptible to forest fires using Machine Learning parameters using various weather parameters and Fire Weather Index(FWI) parameters.

Then if the forest is found to be in danger of forest fires we suggest several precautionary measures which can be taken.Now, to deal with forest fires we present a forest fire simulator. The forest simulator is capable of simulating the real forest fire and help the fire fighters and forest rangers fight the fire more effectively and in a planned manner.

The simulator takes in the wind speed, humidity, temperature, elevation and several other parameters to predict the behaviour of the fire. It would enable the fire fighters know ahead of time where the fire would spread. Also, the simulator has the ability to visualise the fire-line. It can show the behaviour of fire when fire-line is implemented.

For post forest fire recovery, to reclaim the forest land we have presented a deep learning and machine learning assistive tool. This tool using the image of soil and other parameters like pH, drainage, texture can provide the best and effective way to manage the soil, so that the forest land can be conserved.

How we built it

  • The front-end website was made using html, CSS and JavaScript deployed on GitHub pages.
  • The Machine Learning and deep learning models were coded in python and Streamlit and deployed on Heroku.
  • The simulator was made using Unity.

Challenges we ran into

  • Integrating the Machine Learning model was very difficult as we faced a lot of problem.
  • Faced a lot of problem while deploying on Heroku because of the large ML models.
  • Due to time constraints we had technical issues uploading and editing our project.

Accomplishments that we are proud of

  • Successfully able to implement all the ideas we thought of!!!
  • Achieved our three fold motto of protection, prevention, and restoration.

What we learned

  • Various Ml algorithms
  • Integrating ML with web
  • Deployment on Heroku.

What's next for Faunus

Improving upon the recommendation systems and simulator by providing them with more data and actual real world visualisation.

Share this project: