Over the past 18 months, more and more of my energy industry clients have been hiring data scientists, opening innovation centers focused on machine learning, and partnering with artificial intelligence-focused startups. This is no surprise as oil patch data streams have continued to explode from multiple surface and subsurface sources: 3D seismic surveys, real-time formation evaluation, overhead imagery from drones and field survey markings. This is creating unique big data challenges – from processing and analyzing mass amounts of data to matching sub-surface data with accurate surface locations.
Given these trends, I set out to determine the biggest potential for machine learning to help solve these big data challenges in oil and gas. My research led me to an insightful interview by Susan Nash, American Association of Petroleum Geologists, with Paradigm’s Kamal Hami-Eddine, Breakthrough Big Data and Deep Learning in Today’s Oil Industry, which revealed an exciting opportunity for DigitalGlobe. When asked what the most exciting applications for big data were, Kamal answered:
I am biased, and big data cannot go without deep learning for me. For daily usage, the autonomous vehicles development is to me the application which will have the biggest impact on the non-digital world. However the ones which look really impressive to me, are the ones dealing with the live treatment of satellite images. This can stack up to several tera-bytes a day, to transfer, process and analyze.
Here at DigitalGlobe, I’ve witnessed how machine learning paired with satellite imagery has unlocked new applications across industries: population-density mapping, road and structure identification and coastline change detection. And as the leader in high-resolution satellite imagery and cloud-based geospatial big data computing, DigitalGlobe has been harnessing machine learning to innovate alongside our oil and gas customers to drive efficiency in their operations.
How is DigitalGlobe helping O&G customers adapt? Our combination of high-resolution satellite imagery, advanced machine learning, and powerful cloud computing enables feature extraction to be done in days rather than months – helping customers optimize production costs by automating time-consuming field surveys, data input, and analysis.
As you begin planning for 2018, take a look at how DigitalGlobe’s best-in class geospatial data, automated building footprint extraction, and high-resolution digital elevation models are helping customers make more timelyand accurate decisions.
Machine Learning: Building Footprints at Scale
DigitalGlobe, in partnership with Ecopia, has introduced the first (and only) global building footprint extraction product that employs advanced high-resolution satellite imagery, artificial intelligence and cloud compute power. Access to highly accurate building footprints at a global scale has enabled upstream customers to optimize well pad positioning and midstream customers to comply with HCA regulations.
Until now, identifying building footprints has been done manually by a GIS analyst tracing imagery and by survey crews conducting field verification. DigitalGlobe combines cloud computing with advanced machine learning algorithms to create and verify highly accurate building footprints over any area of interest. Our solution offers initial baselining and enables change detection, allowing in-house GIS analysts more time to do what they are trained to do: analyze.
As a result of this capability, our customers are saving on surveying costs, making more accurate well pad placement decisions to ensure regulatory compliance, and improving EH&S outcomes by mitigating risks.
Big Data: High-Resolution Elevation Models
When customers plan new exploration sites, they require high-fidelity elevation models to determine access roads, optimize cut and fill for well pads, navigate residential areas and avoid floodplains. DigitalGlobe now offers a suite of Digital Elevation Models (DEMs), including 5 m resolution DEMs for low-cost, rapid delivery, and global coverage—as well as the ability to rapidly produce up to 50 cm DEMs when you need the highest resolution available.
Using DigitalGlobe’s Geospatial Big Data (GBDX) platform, which houses our imagery library in the cloud, multi-view photogrammetry, and advanced analytics and machine learning capabilities at scale, we’ve expedited the DEM creation process to drive faster results at a fraction of the cost needed for aerial LiDAR or in-field surveys.
Discover how DigitalGlobe’s machine learning-enabled services can drive cost savings in your workflows and mitigate risk.