New AI model detects and categorises marine plastic
A new artificial intelligence (AI) model, able to recognise and classify different types of marine plastic in images shot by a video camera, has been pioneered by a team of scientists at Plymouth Marine Laboratory (PML).
Using a ‘low-cost camera’ mounted on the side of a boat, the model is able to categorise the presence or absence of plastic in an image with an accuracy of 95 per cent. It is also capable of differentiating different types of plastic, for instance, a plastic bag or bottle, with an accuracy of 68 per cent.
Researchers Sophie Armitage and Dr Dan Clewley at PML told Resource that the model was initially trained to be able to recognise three types of plastics: bags, plastic bottles and buoys. These classes of plastics were chosen by the scientists as they are commonly found, but the model can now be trained to recognise other types of debris.
The AI ‘works best when plastic is floating on the surface’, although PML says that the AI still recognised a plastic bag when ‘partially submerged’.
Footage is run through PML’s ‘high performance computer’ – the MAGEO supercomputer (Massive GPU Cluster for Earth Observation), which is based at the lab and operated by the Natural Environment Research Council Earth Observation Data Acquisition and Analysis Service (NEODAAS).
The low cost nature of the camera, PML says, means that the method can easily be used worldwide by boat owners, ‘citizen scientists’, and other researchers.
Different types of plastic often come from different sources, behaving differently in the water in terms of how they degrade. Understanding the characteristics of these different types, PML says, helps provide more information to better understand the problem.
Plastic waste plays a large role in the global pollution crisis, affecting marine organisms and ecosystems and threatening human health. Monitoring marine plastic, PML says, assists in the mitigation of plastic, an often challenging task due to the ‘scale, complexity and time required to do so manually’.
Building ‘a better picture’ of where plastic is ending up
The researchers expect the new method to ‘greatly support efforts to clean up our seas’, potentially through the technology’s ability to identify ‘hotspots’ of marine litter and enable an improved understanding of how floating plastics travel in the water. The technology could be attached to crewed or autonomous vessels, such as PML’s proposed long-range autonomous research vessel, the Oceanus.
For now, PML would ‘like to continue refining the algorithm’, collecting data from a wide range of areas and in different conditions. The lab intends on using the information collected to ‘help develop and validate methods for detecting plastic from satellite data’, providing a ‘better picture of where plastic is ending up in the ocean and where it could be coming from’. Ultimately, this could help with clean-up efforts while providing ‘valuable evidence to inform policymakers’.
‘Simplified observations of marine plastic are currently very limited’
Dr Victor Martinez Vicente, a Senior Scientist at PML, commented: “In situ harmonised and simplified observations of floating marine plastic debris are currently very limited in the literature. We have aimed to tackle the scarcity of these observations through our research on low-cost automated observations.
“We hope that this initial step will lead to an increase of in situ observations everywhere, but especially in poorer countries where marine litter is usually a great problem.
“With the increase of these observations, we expect to support the validation of algorithms from current sensors and the development of future satellite missions. Properly validated satellite algorithms will allow us to use remote sensing techniques to monitor the progress towards Sustainable Development Goals (in particular index SDG 14.1.1.b) at [a] global scale.
“This work is very important to me as I am eager to assist in the battle to mitigate and prevent the devastating effects of marine plastic pollution on our ecosystems.”