CoralWatch

CoralWatch
Home Page
Statistics/Analysis
World Map

Inspiration

We love the ocean, but coral reefs are in big trouble right now because of climate change and pollution. Scientists know they need to keep an eye on the reefs, but they can't be everywhere at once—the ocean is just too big! At the same time, thousands of scuba divers take pictures of reefs every single day. We wanted to build something that turns those everyday divers into "citizen scientists," giving them a way to use their photos to help protect the oceans they love.

What it does

CoralWatch is a web app that acts like an early-warning health monitor for the ocean. Scuba divers and snorkelers can upload the underwater photos they take on their trips. Then, our AI goes to work. It doesn't just look at the photo; it also pulls in real-time satellite data (like ocean temperatures) for that exact spot. By combining the picture with the temperature data, our app can spot early signs of coral bleaching or disease long before it's too late to save the reef.

How we built it

We built the user interface so it's super easy for divers to upload photos and see a map of the reefs. For the "brain" of the app, we used a special type of Artificial Intelligence called "multimodal AI." This means the AI is trained to understand different types of information at the exact same time—in our case, looking at underwater images while also reading numbers like temperature and GPS coordinates. We connected everything using a cloud database so that marine researchers anywhere in the world can see the health scores instantly.

Challenges we ran into

Our biggest challenge was dealing with underwater photos. Water makes everything look blue, dark, and blurry, which can easily confuse a normal AI. We had to figure out how to filter the images so the AI could clearly see the actual colors of the coral. It was also really tricky to get our code to perfectly match a user's uploaded photo with the exact satellite data for that specific spot in the ocean.

Accomplishments that we're proud of

We are incredibly proud that we got the "multimodal" part of our AI working! It's amazing to see the app successfully take a basic photo and combine it with complex satellite numbers to give a real, accurate health score. But honestly, we are most proud of building a project that isn't just a cool tech demo—it’s something that could actually make a real-world difference in fighting climate change.

What we learned

We learned a ton about marine biology and what actually causes a coral reef to get sick. On the technology side, we learned how to connect totally different data sources (like pictures and numbers) so an AI model can process them together. We also learned how important it is to design an app that is simple to use, whether the person logging in is a professional marine biologist or just a weekend scuba diver.

What's next for CoralWatch

We want to add fun features, like a leaderboard and digital badges for divers who upload the most helpful photos. We also want to build an interactive "Learning Mode" so science teachers can use our app in classrooms to teach students about marine life. Eventually, our big goal is to partner with coastal dive shops and eco-resorts so they can use CoralWatch to monitor and protect their local waters.