CIRC07_RosAthene

Inspiration

Every year, more than 9 million tons of plastic end up in the ocean. This is equivalent to setting five garbage bags full of trash on every foot of coastline around the world. This has a negative impact in costal communities since it damages touristic areas, it harms the ocean ecosystems and has a detrimental effect on the health of the people who get exposed to microplastics. For this reason, several organizations have made it their duty to organize ocean-cleanups and dedicate efforts to remediate the pollution problem. However, the fast rate at which the pollution is increasing makes it unfeasible to handle for NGOs on their own.

The world is rapidly moving to a technological era where big data is readily available. That is the case of satellite-data allowing multi spectral analyses with a relatively high frequency, worldwide. Several academics have embarked in the task of using the data to detect plastic debris in the ocean, however, due to a lack of in situ data and the relatively novelty of the field, these efforts are yet to reach their full potential.

What it does

Our solution is twofold. On the one hand, we incentivize the people that live in coastal areas to identify plastic accumulations in their communities and upload the information into our platform, thus providing RosAthene with manually tagged in situ data. This data is then added to a database that includes data extracted from open-source satellite imagery and will allow us to ground the problem in a machine learning optimization problem. By measuring the reflectance observed in the satellite imagery in the areas where local people have confirmed the presence of plastic, we can build a set of algorithms that will help us to understand the pollution routes and cycles so that targeted solutions could be implemented by NGOs to precisely attack the problem.

How we built it

We validated the idea with several mentors from the code4green hackathon in order to make the implementation feasible. Furthermore, we conducted an interview with a member of an Indonesian NGO who provided us with a use case and a deep understanding of the need for such a solution and the impact that it would have to their communities. Finally, we extracted images from several bands from the area of interest, the Citarum river, utilizing Google Earth engine to showcase our solution.

Challenges we ran into

One of the biggest problems in the field of remote sensing is the lack of in situ data. This presented a strong limitation for several approaches that we wanted to take. This makes some techniques such as deep learning and machine learning extremely hard to implement.

Accomplishments that I'm proud of

We are proud of the solution that we came up with. We managed to overcome one of the biggest roadblocks of the project which is the lack of in situ data, by getting the community involved and allowing the locals to participate in our solution. Furthermore, we target a realistic solution which may benefit the remote sensing community as a whole.

What we learned

One of the biggest lessons during this hackathon was that people are willing to help, but a lot of them do not know how to do it. If you provide the community with an opportunity to make an impact, chances are they are going to take it.

What's next for CIRC07_RosAthene

In RosAthene we like to think big. For that reason the team will continue performing interviews with the appropriate stakeholders to better fit their needs. More over, we believe that this is an idea of impact, so we plan to join the SOA ocean accelerator to surround ourselves with experts in the field, and NatGeo's plastic innovation challenge. In 5 years time, we see ourselves as a promising player in fighting plastic pollution with a strong and engaged community of people tagging images and helping fight pollution in Indonesia and some other affected countries.