Leveraging AI and Data from Satellites for Sustainable Development

Advancements in space technologies and sustainable development on Earth are generally considered to be two different perspectives on the spectrum. In fact, there is even a longstanding debate that humanity should focus more on sustainable development on Earth before deciding to focus on space technologies and establishing colonies on Mars. However, there is generally limited awareness of the impact of space technologies on life and development on Earth. The 2030 Agenda for Sustainable Development and its 17 Sustainable Development Goals (SDG) adopted by world leaders at the United Nations Sustainable Development. Data gathered from space contributes to the progress of almost all 17 goals and is expected to play a crucial role in determining the progress over the next decade. How does space contribute to sustainable development? And what is the role of AI?

Ever since the Apollo 8 astronaut William Anders took a picture of the Earth while orbiting the moon (the photo famously called “Earthrise”), both scientists and the general public have been curious to learn more about our “pale blue dot”, which led to launch of various satellites to monitor the planet. Satellites have been playing a crucial role in observing the Earth and protecting our lives – from predicting floods to supporting disaster relief operations, from tracking the melting of glaciers to detecting the spread of wildfires and more – for over three decades. However, recent advancements in space technologies, including the decreasing costs of launching satellites and the increasing miniaturization of satellites and sensors, mean that there is so much more data expected to be available about the planet, than ever before in the history of humankind. From emission data to observe the extent of greenhouse gases in the atmosphere to infrared red of wildfires around the world, there is expected to be petabytes of data available in the upcoming years. This sub-segment of the space industry, called Earth Observation, is expected to become a multi-billion market in the upcoming decade, thanks to the existing and planned constellation of satellites, collecting data about the Earth in real-time. More importantly, this phenomenon is happening at a time where two different, but very related technologies are peaking in their advancement cycles – cloud computing and artificial intelligence.

In the past, any application of data from satellites – tracking glaciers, detecting wildfires, or observing coral reefs – were not automated. Earth observation (EO) was a domain mostly limited to scientists and the experts in the field of remote sensing, simply because they possessed the knowledge to identify objects on a satellite image. However, the recent advancements in cloud computing and artificial intelligence, have meant that EO data can now become truly democratized – meaning anybody without an understanding of remote sensing science, can start to make sense of data from satellites, capitalizing on the lowering costs of cloud computing. Instead of relying on experts for the processing of satellite data, non-specialists can start making sense of satellite imagery. Governments now have the possibility to use AI models on satellite imagery to be alerted of illegal deforestation, including the species of trees being cut and the impact of the deforestation on carbon emissions. Companies now have the possibility to leverage on AI to detect emissions from their manufacturing plants or data centers in order to determine their emission reduction goals. Citizens will soon have the possibility to receive alerts in case they are in a vulnerable zone for wildfires, powered by a combination of machine learning algorithms and satellite data. These are just some of the applications of EO, with plenty more expected in the upcoming years.

Data from satellites provide a unique perspective of the planet due to both the historical availability of data (running up to a few decades from today) as well as the geographical spread. Satellites today can take a picture of every single point of the Earth every day and soon, there might be hourly data of the planet available for us. This is a huge advancement in the context of both monitoring the present and predicting the future. The applications of this technology cannot be seen more so than in the case of monitoring sustainable development across the world. EO data is expected to play a part in both policy creation and policy monitoring in the context of the UN SDGs. The following table provides a summary of the goals and the respective indicators that EO data plays a direct or indirect role.

For instance, night lights data collected from satellites are being used in developing economic models by institutions such as the World Bank and the International Monetary Fund, to track the decrease in poverty in the less developed countries of the world, using electrification as a proxy. Using artificial intelligence models to automate the identification of lights over an area and classifying them to factories and residential complexes, this approach helps monitor SDG 1: No poverty. Similarly, satellite data is also used by farmers around the world to provide advisory services on irrigation, fertilizer management and crop production. In Africa, the economic impact of using EO data in agriculture has been estimated as almost $1 billion by 2024, thanks to the automation of applications through AI. This corresponds with working towards increased food security for African citizens contributing to SDG2: Zero Hunger. Finally, as an example for SDG14: Life below water, communities around the world use EO data to monitor water quality and assess the conditions and impact on marine life to improve inshore management, leveraging on the ability of machine learning models to detect algae blooms, pollutants, and plastics under the surface of the ocean. With increasing data expected to become available over the next few years, along with an increased resolution of imagery (<30 cm), the computer vision models to detect such objects are going to become more and more efficient, resulting in an automated monitoring service.

Historically, we used to look at space for inspiration, looking at the planets, the stars and the constellations. But currently, we now have the ability to look down upon our planet for insights about our development. However, unlike the insights in the past few decades, this decade we have the ability to automate the monitoring of our planet and make the observations more efficient and actionable, thanks to the advancements in artificial intelligence and computer vision. With an existential crisis looming in the name of climate change, along with the increasing need to focus on a sustainable future, space and more specifically, data from satellites have a huge role to play as supporting tools in determining the future of humanity – to help us to monitor our actions, to provide updates on our activities as well as to predict the evolution of our biodiversity and environment. It is thus, necessary, to integrate the use of satellite data and space technologies and to explore the various possible use cases thanks to developments in AI, as we look towards re-developing our economies after the pandemic, towards a sustainable future.

References: