The SMART project, led by CERENA researchers, won the 1st edition of the AI Moonshot Challenge and will receive €500,000 to develop a solution that will help to fight plastic pollution in the Oceans. The announcement was made this Thursday, 3rd December, by the Portuguese Space Agency (Portugal Space), at the Web Summit.

Using satellite data from the Copernicus program, the team intends to determine which frequencies are appropriate for the detection of plastic in bodies of water. “Our idea is to develop a tool that creates a map showing the probability of plastic pollution in our Oceans, in the past, in the present and in the future”, says Leonardo Azevedo, Técnico professor and CERENA researcher. “We want to make these maps available to the authorities, which will then define and develop strategies to reduce and mitigate this problem”, adds the professor.

Using Data Science methods and Artificial Intelligence (AI), the team will detect floating plastics from satellite images, by combining them with high-resolution numerical models and data science methods. The prediction model will be validated in loco using autonomous underwater robots. “Through this validation or absence of validation, it will be possible, if necessary, to update the model and improve the probability maps”, says professor Leonardo Azevedo.

According to Portugal Space, the jury also valued the fact that the consortium’s partner network plans for the use of autonomous marine vehicles to validate the results and collect even more data”

The CERENA researcher also highlights the relevance of this method to estimate the accumulation of plastics in the ocean over time: “It is very important because it allows us to know how did plastic get into the Ocean. One of the biggest problems we have is not knowing when and where plastic gets into the Ocean. This is a critical issue because the longer plastic will remain in the Ocean, the more it will break down into millions of tiny plastic particles, which will end up being ingested by fish, thus entering the food chain”.

“We are all very happy to be involved in a project that adds real value to our society and that responds to a global problem that affects us now and will do so in the future”, said the CERENA researcher.

The project team is composed of professor Leonardo Azevedo (project leader), professor Amílcar Soares and professor Maria João Pereira, who are also Técnico professors and CERENA researchers, professor Zita Martins, from the Department of Chemical Engineering (DEQ) and CQE researcher. Experts from the Laboratório de Sistemas e Tecnologia Subaquática (FEUP), the Centre for Environmental and Marine Studies (CESAM), the Hydrographic Institute and MIT-Portugal, are also part of the team.

“This is a very multidisciplinary team and the project brings together experts in Spatial Data Sciences from CERENA, experts from FEUP and University of Aveiro in Marine Sciences and Oceanography, with a lot of knowledge in underwater robotics”, says prfessor Leonardo Azevedo. The project also has the expertise and skills of MIT researchers in AI and Machine Learning. “The Hydrographic Institute plays an important role in validating our models”, he adds.

According to professor Leonardo Azevedo, the combination of this multidisciplinar knowledge was the most prevalent criteria in the evaluation of the applications. “We had a multidisciplinary approach, from a new perspective. This vision has given us some advantage, and we are all very young, which is good. It is yet another example of the AI capabilities in Portugal. This project also demonstrates the quality of research carried out in Portugal in this area, particularly at Técnico and CERENA”.

The SMART project won a €500,000 cash prize for development and implementation. The project will last for two years.

Researchers from around the world set out to find solutions to the problem of plastics in the Oceans

Ten projects from 13 countries applied to the 1st Edition of the AI Moonshot Challenge, launched in 2019 at the Web Summit. This international challenge attracted researchers from the United Kingdom, Serbia, Estonia, Finland, Brazil, Canada and the United States, among many others. In total, 38 institutions joined together in ten consortia, proposing to find solutions to the problem of plastics in the Oceans through the use of satellite data and AI.

The AI Moonshot Challenge, promoted by the Portuguese Space Agency, Portugal Space, in cooperation with Fundação para a Ciência e a Tecnologia (FCT), Unbabel, European Space Agency (ESA), the National Innovation Agency (ANI) and Web Summit, is an international competition that aims to find disruptive ideas combining satellite data and Artificial Intelligence for the detection of plastics in the Oceans.

The international jury was co-chaired by Paolo Corradi, systems engineer at ESA and biologist Carolina Sá, project manager at Portugal Space. Nikolai Maximenko, researcher at the University of Hawaii (USA), Laura Lorenzoni, NASA scientist (USA), Vasco Pedro, CEO and co-founder of Unbabel (Portugal), Shungu Garaba, researcher at the University of Oldenburg (Germany) and Pierre-Philippe Mathieu, responsible for the Explore Office at Φ-lab from ESA), were also members of the jury.