There is a huge market of illegal fishing. The UN estimates that 1800+ pounds of illegal fish are caught every second in the world! The global market of illegal fishing is $23.5 billion.

Illegal fishing encompasses a lot of problems - but at the root of it all - it gravely affects the fish in our oceans. A major cause of Illegal fishing is lack of proper information passing and sharing between fishermen and local authorities. This lack of information is in the form of ambiguous laws and blurry boundary lines for fishing areas. This is where SeaSure finds its niche.

What it does

SeaSure is an app that tries to deter illegal fishing by alerting local authorities every-time a fishing law violation in the form of fishing in a restricted/ out-of-bounds area is made. It maps fishing routes, pinpoints legal and illegal fishing spots, and detects a fishing vessel's compliance with its local fishing laws.

SeaSure is reliable, cheap to deploy, scalable, and most importantly, can work offline due to its heavy usage of the users' phone GPS coordinates.

How I built it

The app uses a fisherman's phone GPS coordinates to determine the location of their fishing vessel. It then runs an algorithm to determine the speed of the vessel. Based on the vessel's speed, the ML algorithm determines if the vessel is sailing in the waters or catching fish. Subsequently, SeaSure drops markers at every fishing spot that it detects.

In the backend: A Selenium program scrapes the web for fishing laws and regulations based on specific locations acquired by the phone GPS coordinates. A Recurrent CNN model classifies the large amounts of text received into a database. Based on all the information collected till now, python is used to make fishing boundaries. Any fishing spot marked within the boundary is marked green. Anything beyond is marked red. Every red fishing spot gets reported to the local authorities.

The alert system for local authorities is also based on the GPS coordinates collected through the user's phone. The pin-code that corresponds to the local area is first queried. Then, the fishing authorities associated with that pin-code are alerted vis SMS. The SMS contains the type of violation, as well as real time location of the fishing vessel.

More details on technologies used are present in the video!

Challenges I ran into

So many. I was unsure of how to get the Recurrent CNN model working. But I found a few easy to follow notebooks on Kaggle and GitHub that gave me enough understanding on how to begin.

Also finding and collecting reliable datasets was so so hard.

Accomplishments that I'm proud of

Everything :,) This was a really long project and I was doing it all alone. I had to google and lean so many new things as I was going.

What I learned

I learnt a lot about GPS tracking and the importance of data sharing

What's next for SeaSure

  • Implement the Recurrent CNN script the can help classify texts of legal documents.
  • Include a geo-tracking feature that will allow officials to locate the vessel in real time, with better ease.

Built With

Share this project: