Big ‘Earth Observation’ Data: Challenges and Applications

As nearly a thousand Earth observation satellites currently orbit the planet, terabytes of remote sensing data and satellite imagery of land, vegetation, water bodies, glaciers, urban landscapes, and other geographic features become available for end users across multiple industries.

Modern GIS systems allow the collection of all such geospatial data in one place for a comprehensive analysis of the area under study.

The arrival of Big Data set the stage for more advanced processing of EO datasets and garner additional insights through machine learning, predictive modeling, and other cutting-edge applications.  

Big Data technology enables users to manage and process vast amounts of versatile data (for example, satellite pictures of Earth), which is getting bigger by the day.

The convergence of GIS and Big Data was a matter of time and tech progress, as their union opens up immense opportunities for businesses and individuals alike.

Many organizations are gaining value with geospatial big data analytics to tackle existing issues and improve decision-making and future planning.

Ten years ago, the high price of hardware computing resources was the main barrier that kept many from using Big Data. Today we see more affordable hardware costs and have access to scalable cloud products, which altogether makes this technology available to anyone, from large governmental institutions to tiny private businesses. The present challenge is about talents; not everyone has the expertise required to leverage big chunks of data to the benefit of their companies.  

Satellite images of Earth have proved their unrivaled value in multiple domains, including agriculture, forestry, environment, disaster relief, intelligence, defense, and urban planning among dozens of others.

Here’s looking at some less-known industries that are leveraging the ability of remote sensing to reveal what the human eye can’t see. 

Over the past years, satellite images of Earth have become frequent in the news media providing a deeper context for the tragedies occurring on the ground. With the help of remote sensing satellite data, not only can we see the consequences of destructive natural or manmade disasters and pinpoint where illegal forest clearing or fishery is made, but also get clues about what or who is responsible for the damage done in war conflicts, and even find better solutions for mitigating the impacts.

Bellingcat is one good example of how a group of journalists can make some eye-opening discoveries about international conflicts based on freely available geospatial data. Through the mediation of Earth observation specialists, journalists can access one more source of relevant and up-to-date information for their stories, which is satellite imagery.

Geospatial data has been widely used to monitor various environmental issues, such as deforestation, melting glaciers, water bodies polluted by oil spills, and so on. No wonder there are known applications in air pollution studies as well.

As we know, agriculture accounts for the largest portion of harmful NH3 (ammonia) emissions, which are only expected to rise with more ammonia plants set for opening across Asian and Middle East countries by 2030. The slurry that is stored in tanks until it’s ready for application as fertilizer is the main source of ammonia.

One of the recent use cases is the collaboration of the Danish government and a satellite data provider to estimate the country’s total NH3 emissions. The combination of machine learning and the most up-to-date satellite images of farms helped to detect 26,000 slurry tanks in Denmark within hours. With internal data on each farm and its production type at hand, they were able to produce a countrywide estimate of ammonia emissions coming from slurry tanks.

Satellite pictures of Earth captured by multispectral and hyperspectral sensors can benefit exploration geologists and scientists too. At the very least, geologists can gather reflection data and absorption properties of various soil types, rocks, and vegetation cover to differentiate between the Earth’s surface features. 

On a more advanced level, certain spectral bands can be used at several stages of mineral mapping and exploration. Mining companies that rely on spectral analysis of satellite data save time and costs as it helps narrow down the search and map the potential areas with mineral deposits.

The current amount of GIS data in our hands is overwhelming. It takes the form of remote sensing imagery, real-time satellite view feeds, topographic maps, LiDAR and inSAR data, and others, providing a myriad of opportunities for use.

Knowing one’s way around huge datasets and being able to extract value for each particular industry with existing solutions presents one of the main challenges for modern GIS specialists.