By Amy Challen, General Manager Artificial Intelligence on Jul 6, 2022
Can computer vision techniques help assess the health of ecosystems?
Looking at social media photos might seem an unlikely way to start on an environmental conservation project, but a Shell GameChanger project in Brazil did just that. UK-based start-up Space Intelligence Ltd. uses geotagged images to improve the interpretation of satellite images of vegetation and ground sensor data. Their successful proof of concept (POC) demonstrated that artificial intelligence technology (AI) can help assess how well ecosystems serve as carbon sinks and can predict how they might change over time.
With computer vision technology, AI is used to interpret and understand the visual world by leveraging untapped data, in this case the images willingly shared by users of social media networks. By complementing on-site inspections, Shell believes these additional insights can increase trust in the efficacy of nature-based solutions (NBS) projects.
Nature restoration projects are part of the solution
According to the United Nations Intergovernmental Panel on Climate Change1, protecting and restoring natural ecosystems can help to limit global warming and bring additional environmental and social benefits to local communities, including job creation and improved biodiversity. Indeed, that is what NBS projects aim to do: to protect, transform or restore a habitat, thereby helping nature absorb more carbon dioxide (CO2) from the atmosphere. Such activities can lead to the marketing, trading and sale of high-quality nature-based carbon credits.
Shell aims to invest around $100 million a year in nature-based projects that reduce atmospheric CO2 levels or avoid CO2 emissions in support of our climate target. These projects also deliver other benefits such as improvements in biodiversity, water quality, flood protection and livelihoods for people in local communities.
NBS have a role to play in reducing the impact of the carbon emissions from the energy products that we sell, by offsetting the unavoidable emissions associated with a customer’s use of our energy products. We believe that they are part of the solution, not the solution to climate change – and must only ever be used in addition to robust decarbonisation across the global economy.
Evaluating the contribution that NBS projects make to carbon offsetting can be challenging, however. How can we ensure a project helps establish or maintain stronger biodiversity and increases an ecosystem’s capacity to absorb CO2? Analysing satellite images is a crucial part of this process, but it is not always easy to reconcile the images taken from orbit with the truth on the ground.
Proving the commercial viability of AI for NBS
That may soon change, thanks to the technology developed by Space Intelligence. Their POC sought to map the past and present vegetation of parts of the Cerrado, a vast tropical savanna ecoregion in Brazil. The Rebrota project (rebrota is Portuguese for regrowth) calibrated satellite imagery against geotagged pictures of plants and animals that people in the study area have uploaded onto social networks. This technical solution was chosen for further development from more than 70 proposals submitted in response to a Shell GameChanger open Call for Solutions.
We're grateful to Shell's GameChanger programme for enabling a highly innovative project developing technologies which provide actionable intelligence to support quality NBS.
Space Intelligence uses a special algorithm to smartly assess any changes in the natural environment that might impact its ability to absorb CO2. The tool then combines these insights with geospatially located images from satellites to help provide a broader picture. In short, they can monitor biodiversity and predict the evolution of parts of the Cerrado degraded by failed attempts at farming.
Another step in the right direction
Eventually, the project could help to determine whether these areas are becoming thriving, net-carbon-absorbing ecosystems or degrading into areas with net-carbon emissions. This POC is encouraging. In future, AI-based systems that combine satellite imagery with openly shared pictures of wildlife could be more widely used to evaluate, track and assess ecosystems performances and the value of NBS projects worldwide. This could open the door to more holistic and increasingly transparent monitoring of NBS projects using computer vision technology.
Digital innovation results from rapid prototyping and the ability to quickly identify the solutions that can be deployed at scale. Shell already uses satellite imagery to detect, quantify and avoid harmful emissions in oil and gas operations, such as methane leaks from pipelines. Working with start-ups like Space Intelligence is truly inspiring because of the external perspective they bring to how we approach the dual challenge of the energy transition and the digitalisation of society. I wish them every success in further developing their technology!
To find out more about the work of Space Intelligence, listen to their Chief Technical Officer & Co-founder Professor Ed Mitchard in this webinar: Can AI and digital tech accelerate the energy transition?
1 IPCC, 2019: Summary for Policymakers. In: Climate Change and Land: an IPCC special report on climate change, desertification, land degradation, sustainable land management, food security, and greenhouse gas fluxes in terrestrial ecosystems.