Supporting Marine Conservation with Image and Video Annotation

Feb 16, 2021
“Every year, about 8 million tons of plastic waste escapes into the oceans from coastal nations. That’s the equivalent of setting five garbage bags full of trash on every foot of coastline around the world.”
– National Geographic, The world's plastic pollution crisis explained

The world’s oceans are facing an unprecedented challenge in the twenty-first century. A combination of pressures, all the result of human action, are threatening the sustainability of ocean environments across the world. Plastic waste, overfishing, and climate change can seem to be insurmountable problems, however, recent developments in computer vision based AI technology are helping researchers and environmentalists to better understand and address these issues.

AI for marine conservation is an exciting field that needs support in the form of precise, affordable image and video data annotation. This blog will look at how data annotation services, like Keymakr, can form part of the solution for protecting our threatened oceans.

Spotting plastic waste with AI

The accumulation of plastic waste in the ocean is a significant risk to wildlife and humans alike. Larger pieces of waste can cause damage to wildlife, whilst over time plastic breaks down into microscopic particles that make their way into the food and water supply for all of us. Governments and environmental groups have begun to reckon with the difficult task of collecting plastic in the ocean before it can cause more harm. One of the key challenges is identifying smaller areas of plastic waste across vast areas of ocean. AI models are being developed with the capacity to identify plastic waste in satellite or drone images. This technology allows for a huge amount of data to be analysed, meaning that problematic accumulations of plastic can be spotted and cleaned up.

Spotting plastic waste with AI | Keymakr

The training data required for plastic spotting computer vision applications is specific and difficult to acquire. Crowdsourcing has been proposed as a solution, with members of the public providing images of plastic. However, dataset creation can also be supported by professional annotation services. Experienced annotation providers are often best placed to respond to complex data needs due to specific expertise and access to the right annotation tools.

Finding and recording marine wildlife

AI is also being deployed to monitor and help protect vulnerable ocean wildlife. There are many potential marine conservation applications for computer vision technology. Whales, due to their large size, are visible from satellite imagery. Deep learning algorithms are being developed that can discern the shape of whales amidst a huge quantity of satellite image data. This valuable information will allow conservationists to track individual animals, and assess overall population numbers for threatened species. AI monitoring is also being put to use to make fishing more sustainable. This technology can identify illegal fishing, and transportation of fish. This will make monitoring fishing stocks significantly more efficient.

Video annotation for marine wildlife monitoring | Keymakr

Aerial AI training data contains large amounts of information that must be precisely annotated and labeled. Video data in particular, can be a challenge for smaller charitable or research organisations to accomplish. Professional annotation services are able to leverage bespoke annotation tools to create video training data that supports AI for ocean conservation.

Analysing coral reef health

Coral reefs, perhaps the most diverse and vital of ocean ecosystems, have experienced a shocking decline in the past decade due to coral bleaching and climate change. Efforts are now being made to speed up protection of these fragile environments by using computer vision powered cameras to monitor coral and animal health. These cameras have a much lower impact on wildlife and cause less disruption than traditional monitoring by divers. They can be used to measure the numbers of different fish species or to assess the degree of coral bleaching and degradation.

Spotting life in coral reefs | Keymakr

These vital advances in monitoring technology can be facilitated by smart annotation and quality control. In-house annotation, carried out by managed teams of annotators, can provide an edge to AI development. Keymakr can ensure that today's innovators have access to the right data, at the right scales. Contact a team member to book your personalized demo today.

pixel-perfect image and video annotation
Great! You've successfully subscribed.
Great! Next, complete checkout for full access.
Welcome back! You've successfully signed in.
Success! Your account is fully activated, you now have access to all content.