Using AI to Identify Plastic Pollution and Expedite Ocean Cleanup

Oceans cover 71% of our planet and contain a volume of plastic that has become so high that “in just a few years, we might end up with a pound of plastic for every three pounds of fish in the sea” (1). Every year, more than 8 millions metric tons of plastics enter our oceans on top of an estimated 150 million metric tons already circulating through them (1). 8 million, an amount equivalent to “dumping one New York City garbage truck full of plastic into the ocean every minute of every day for an entire year” (1). The Pacific Garbage Patch has become a poignant icon of ocean pollution across media outlets, environmental campaigns, and social media platforms. We’ve all seen the images of birds nesting in piles of garbage, sea turtles with plastic straws up their noses, and fish entangled in netting that are displayed across the internet. But even more horrifying is that “approximately half of all plastic pollution is submerged below the ocean surface, much of it in the form of microplastics so small that we may never be able to clean them up completely” (2). There is a strong demand from governments, environmentalists, and citizens alike to keep our oceans clean and preserve the biodiversity and critical services they provide to life on Earth. Many emerging companies and organizations are changing the landscape of ocean conservation by developing innovative AI monitoring techniques and robotic technologies to identify pollution in our oceans and waterways and provide an expedited approach to cleaning it up. AI and machine learning enables autonomous ocean vehicles to distinguish plastic from other objects and materials, allows for the identification of large concentrations of garbage, and helps tackle pollution at the source.

“It is essential to identify which parts of the ocean collect the most plastic to effectively target cleanup and pollution prevention efforts.”
— Liz Allen, Environmental Regulatory Specialist, Forbes

Though the world has been producing plastic and thus, waste, since the 1950s, “it was not until the 2010s that governments and other international bodies really began to address the problem of plastic pollution” (13). According to a UN report, there was a huge spike in laws to limit single-use plastics in 2014, with limited regulations before mainly due to the fact that “many people were unaware of the scale of the problem” (13). This shift in public awareness can be attributed to increased exposure to plastic pollution and waste imagery with the rise in social media (13). Now, with a growing public awareness, rise in social media imagery, and increased regulatory pressure and law, people are realizing the full extent of the damage being done to our oceans, requiring ongoing monitoring of the scale of the problem and the progress being made.  

How AI is Changing Ocean Cleanup

By combining machine learning, satellite data, and drone technology, scientists are equipped to remotely detect plastic waste hotspots, quantify marine litter, and even detect tiny plastic pieces floating in oceans. (11, 13). 

In the past, ocean cleanup missions have been restricted to shorelines and coastal waters due to a limited human capability to reach and collect plastic circulating far out at sea. Plus, cleaning up large patches of ocean debris can be a tedious and labor intensive process. Current methods of tracking debris distribution in the ocean typically “involve surveying the damage from planes and boats” (15). Detecting plastics in the ocean is “notoriously challenging”, being that plastics come in many colors, sizes, and shapes, and are made from “a variety of chemicals” (11). Recently a sunken container ship off the coast of Sri Lanka has led to the spillage of 87 containers full of plastic pellets, called nurdles (12). Since the spill in May 2021, nurdles are still washing up in the billions along Sri Lanka’s coastline, as well as being found in the bodies of washed up dolphins, turtles, and fish (12). Nurdles are the pre-production plastic pellets used as the building block for all our other plastic products, and can be made of a variety of plastic chemicals (12). The cleanup continues today in Sri Lanka, mostly by cleanup crews on shore.

Satellites typically offer the best way to detect objects floating on the ocean surface, but as is the case of nurdles, plastic debris is often too small for their resolution to detect (14). To tackle this issue, researchers in the UK have for the first time used satellite data to detect patches of tiny plastic pieces down to 5 millimeters in size (14). Scientists at the Plymouth Marine Laboratory used the combined efforts of two European Space Agency Sentinel-2 Earth Observation satellites equipped with 12-band Multi-Spectral instrument (MSI) sensors to repeatedly collect data “from all coastal locations around the world every 2 to 5 days” (11). 

“Materials can be distinguished using light signals based on which wavelengths of light they reflect” (11). 

While water absorbs light in the near-infrared to shortwave infrared light range, floating materials like plastic and debris reflect in the near-infrared (NIR) instead (11). Using the satellite data, researchers trained a machine-learning algorithm to analyze the images pixel by pixel and associate certain NIR light signals with floating plastic debris (11). “They also taught the algorithm to distinguish between plastic and natural materials such as seaweed, driftwood, and seafoam” (11). When put to the test in four coastal waters around the world, the algorithm detected plastic with 86% accuracy across the four locations, with an accuracy of 100 percent off the Gulf islands in Canada (11).

Researchers at the University of Barcelona have developed an aerial imagery algorithm for detecting marine litter. The team trained the algorithm to analyze 3,800 aerial images of the Mediterranean and used neural networks to improve its accuracy over time, winding up with a reliable artificial intelligence tool to detect and quantify plastic on the surface (15). The tool “can analyze images individually or sort them into segments...to offer an estimate of density” (15).  Team member Odei Garcia-Garin shared that the new algorithm “reaches a 80 percent precision in the remote sensing of floating marine macro-litter” (15). The tool is an “open-access web app available to professionals in the field, and the team expects to develop a version that can work with drones, to fully automate the process” (15).

AI and machine learning algorithms can be applied to drone imagery and is being used widely in the tracking and subsequent clean-up of plastic pollution. The director of Plastic Tide started “with the idea to use drone-mounted cameras to take thousands of aerial photos” which are then used to train an AI algorithm to recognize and distinguish plastic from marine life and other materials (13). These drones are making open source maps of the plastic waste in our oceans to detect hot spots and direct clean-up efforts (13). The long term goal of Plastic Tide is to build a system which can document the spread of plastics in close to real time and provide a way to assess the impact of policies - for example a ban on single-use plastics (13). 

Icon;

Tackling Plastic Pollution at the Source

To locate pollution sources and help prevent plastic from ever reaching the ocean, German masters-student Marcella Hansch set out to find a solution under the analogy that “when your basement has flooded, you want to turn off the tap first before you mop up the floor” (6). Together with biologist Tilman Peter Flohr and journalist Clemens Feigl, she founded the company everwave in 2018, which has developed and uses specialized boats, the so called CollectiX, with a conveyor ramp that can skim trash from rivers, lakes, and dams (6).

When Hansch’s home state in Northern Germany experienced catastrophic flooding in July 2021, “she quickly deployed one of her boats to help clear a nearby river of wood and debris” (6). The boats are equipped with cameras, a drone, and artificial intelligence so that collected trash can be analyzed by the German Research Center for Artificial Intelligence, one of the world's largest nonprofit contract research institutes for AI technology” (6). With the help of a team of engineers, biologists, and environmental scientists, the technology is employed with a goal of detecting what can be recycled and repurposed and identifying the “biggest entry points and who the biggest polluters are'' (6). Many of the deployment missions have been sponsored by donors such as the environmental foundation of automaker Audi, for a mission in the German Danube that yielded more than 7,000 pounds of pure plastic. As a small, donation-based organization of 15 staff members,

Everwave is focusing on scaling up to meet the demand for trash cleanup all over the world (6). Hansch is planning deployments in sub-Saharan Africa and Asia next…

not only because these are areas that leak a large percentage of the worldwide trash into the oceans, but because we export a lot of our trash to these regions” (6).

Ocean cleanup organizations are usually non profit organizations that rely on donations and volunteers. To scale up and incentivize more of it, organizations and foundations like Everwave are looking for investors.

This fall, Everwave began anchoring honeycomb-shaped platforms called HiveX in trash hotspots on rivers and lakes (6). The boats and anchors are able to anchor long-term and are “deliberately designed simply so they can be built locally in Africa, Asia, or Eastern Europe” in order to create jobs for local residents (6). Hansch also hopes that this will incentivise more ocean cleanup missions and technology once “locals start seeing trash as a valuable commodity that can be recycled” (6). 

The Ocean Cleanup is a non-profit organization that is developing advanced technologies to rid the oceans of plastic. Even though everybody wants to market themselves as sustainable, at the end of the day, business prioritizes profit. (6)

The Ocean Cleanup uses machine learning to identify plastic pollution in rivers and simulate how it moves in the ocean. These insights power passive cleanup systems to help remove plastic that impacts our ecosystems. (4)

The Ocean Cleanup has two cleanup solutions - one for rivers and one for oceans - for identifying and cleaning up plastic (4). With this, they are directly acknowledging that the main gateway for plastic entering the ocean is through rivers. Recognizing plastic and distinguishing it from other objects and materials is traditionally no easy task, but by collaborating with a Microsoft Hackathon for Sustainability project, the Ocean Cleanup was able to “build a solution that uses AI to identify and label tens of thousands of images” (4). For plastics in the ocean, a passive cleanup system collects plastics from accumulated patches of debris for extraction (4). Sensors attached to passive cleanup systems also collect data on winds and currents, and “an algorithm runs simulations to see how these cleanup systems move through the ocean” (4). For river monitoring, cameras capture images of debris moving toward the ocean, which are then analyzed to differentiate plastic from organic debris (4). The Interceptor, a solar-powered, autonomous collection unit, then removes the plastic before it reaches the ocean (4). The Ocean Cleanup plans to combine their river and ocean systems at scale “to reduce ocean plastic by 90% by 2040” (4). One of their most recent accomplishments was demonstrating a way of doing this using infrared to distinguish swirling pieces of plastic from other ocean debris (15).  The Ocean Cleanup Project has now “ventured into the Great Pacific Garbage Patch with research vessels and flown over the top of it with aircraft fitted out with sensors and imaging systems” (15). 

Sustainable Coastlines is empowering communities to restore their marine environments by using AI to identify the sources, causes, and solutions of coastal pollution (3). Sustainable Coastlines is a grass-roots level initiative, empowering communities to come together to collect litter from coastlines and then log it in their “uniquely comprehensive database” (5).  Since their start, Sustainable Coastlines has removed “enough trash from shorelines around New Zealand and the Pacific to fill the equivalent of nearly 45 shipping containers” (5). Although an impressive achievement, the problem of plastic accumulation in the ocean is only getting worse, driven by consumerism and a growing global population. With the help of new AI technologies and solutions, Sustainable Coastlines is “developing a national litter database that will not only track the impact of clean-up efforts but also generate accurate, scientifically valid data and insights,” to “guide and help local communities of ‘citizen scientists’ “ (5). 

AI and Robotics to Reduce Ocean Pollution

AI is being utilized in Autonomous Underwater Vehicles (AUVs) and Unmanned Surface Vehicles (USVs) to clean up plastic and other forms of pollution. USVs are robots designed for the efficient and cost effective collection of commercial ocean data on the surface. Similarly, AUVs are robots pre-programmed to collect data from specific parts of the deep ocean while scientists conduct research on board a ship or land.  “Ideally, AUVs should be able to perform complex, long-duration missions without the need for human intervention. This requires state-of-the-art (or beyond) artificial intelligence (AI) techniques” (8). Companies like Liquid Robotics, a Boeing company, are overcoming roadblocks to ocean data by utilizing a network of drone boats gathering ocean data globally. Liquid Robotics designed the Wave Glider, a USV powered by wave and solar energy, that can operate individually or in fleets, delivering a variety of ocean data from the depths to the surface. The Wave Glider is designed to support  a wide range of sensors, allowing “partners to develop and deliver unique solutions to customers” (8). 

Using USVs, XOCEAN provides “turnkey data collection services to surveyors, companies and agencies” from mapping the seabed to environmental monitoring (9). XOCEAN recently adopted Marine AI’s ‘Guardian Vision’ AI software to enhance their USV’s situational awareness (10). Marine AI is a software engineering firm “specializing in edge-based artificial intelligence software” (10). Their goal “is to deliver sensor-driven cognitive artificial intelligence to enhance maritime capability, on or below the water” (10).

Razer partnered with marine waste cleaning startup Clearbot “to advance the use of AI and robotics to reduce ocean pollution” (7).  They announced their partnership in celebration of World Oceans Day in June 2021.  The team designs robots that leverage AI-powered computer vision to identify marine waste and retrieve it to be responsibly disposed of (7). The “Clearbot” is a marine trash collecting drone that’s solar powered, self-navigating, and equipped with vision AI that use machine learning to optimize its capability over time (7). Through the partnership, Clearbot was able to utilize Razor’s technical expertise to optimize their mission and improve the Clearbot design by making it more efficient, marketable, and scalable. “Clearbot is calling on the community to upload photos of marine plastic waste commonly found in open waters to their website that will be used to help improve the robot’s waste detecting AI algorithm” (7).

“With the new model, we’re confident in extending our reach globally to protect marine waters, starting with partners which include marine harbour operators in Asia and NGOs who have already expressed interest”
— Sidhant Gupta, Chief Executive Officer at ClearBot

Conclusions

Our planet’s oceans continue to be threatened by plastic pollution, amongst many other environmental threats. But rather than let our oceans fill with more and more trash, we as individuals can try to consume smartly, manage our waste, and support ocean conservation groups. We have the power to control and limit the amount of garbage that ends up in our oceans and waterways. And as for the millions of tons of waste that already circulates through them, we can leverage innovative AI technology to help clean up our mess.