Marine Litter

Plastic and other marine debris are among the greatest environmental threats of our time. They endanger marine animals, destroy habitats, impair fishing and shipping, and cause massive economic damage in the long term. Studies estimate the global losses caused by marine pollution at several hundred billion dollars per year.

​

Plastics decompose into microplastics, which enter the food chain—posing risks to animals, ecosystems, and ultimately human health. Despite the urgency of the situation, one key problem remains: we know far too little about where marine debris is actually located.

Until now, data collection has been based on local observations and selective surveys. This is expensive, slow, and not scalable. This is exactly where our project comes in.

Our approach

We are developing an AI-supported system that can detect marine debris on satellite images and visualize it in near real time. We use an open-source model to continuously generate new training data: we are already making the data obtained in this way freely available and are using it to gradually improve the model.

The goal is to create a global, freely accessible map that helps organizations worldwide to plan cleanup measures more effectively, accelerate research, and raise public awareness.