But Revealing the Hidden Information Can Be a Challenge
By Natasha Léger


SATELLITE AND AERIAL IMAGERY HAS LONG BEEN THE domain of imagery and GIS professionals. Google Earth changed that perception by literally putting the globe, with the ability to zoom in and zoom out, into the hands of anyone with an Internet connection. However, the information gap between a satellite or aerial image and information or insights is still rather wide for the non-imagery professional.

A satellite or aerial image—a snapshot in time—can be artistically beautiful for anyone to enjoy. It can be scientifically and academically instructive for researchers when combined with change detection over time. It can verify market dynamics such as in agricultural commodities or energy—oil, gas, wind, solar, hydro.

What can satellite or aerial imagery do for marketers, operations directors, business analysts, and other knowledge professionals whose day-to-day workflow and soft- ware tools include an excel spreadsheet, PowerPoint, CRM system, custom in-house software, etc., but not imagery or GIS software?

Location data enables the correlation of seemingly disparate or unrelated information like never before. Satellite and aerial imagery provide overhead views of the world that com- plete the picture developed from other location data sources such as mobile, GPS, surveys, sensor networks, photos, maps, 3D data, addresses, place names and more. It’s another dataset in answering a specific business question.

The same image can answer different questions depending on what other information can be overlaid to tell a meaningful story. This type of data analytics takes a lot of computing power and specialized software. The total cost of ownership—data + storage + software + staff—is generally prohibitive for companies that do not have existing GIS and imagery departments. DigitalGlobe is working to change that.

DigitalGlobe recognizes the power of the information embedded in pixels and has developed a cloud-based platform that enables feature extraction, and the conversion of those features into quantifiable data at scale.

For example:

→ How traffic patterns around a particular intersection will affect the ROI of a billboard campaign;

→ A regional market’s ability to absorb higher food prices due to local, regional, and global drought conditions;

→ How vehicle and truck activity at factories worldwide affect the global supply chain for particular industries; or

→ How the location of buildings across geographic footprints are related to their roof composition and age. 

You can get the ground and overhead perspective to support your business plan or presentation without breaking the bank. But, you will likely need to find an industry market analysis company that integrates vast amounts of data, including satellite and aerial imagery, to deliver your answer. DigitalGlobe is working with a variety of channel partners in Financial Services and Retail, to name a few, to ensure that their end customers—marketing profession- als, operations leaders, business analysts—get the benefits of the insights available from imagery analysis.

An image hides thousands of pieces of information that can answer a business question. Shay Har-Noy, Senior Director, Geospatial Big Data for DigitalGlobe, joined DigitalGlobe as part of their acquisition of Tomnod in April 2013, where he was CEO/Co-founder. Tomnod is a crowdsourcing platform for feature identification from imagery. Fully automated feature extraction remains mostly limited to such things as buildings, cars, airplanes, ships, certain agricultural crops, etc. Har-Noy said, “I am looking forward to working with businesses to leverage automated feature extraction and crowdsourcing to answer their imagery-related business questions.”

DigitalGlobe and other imagery companies around the world are committed to turning pixels into information. With the help of other analytics companies and platforms, as time goes on, satellite imagery, with improved feature extraction and analytical tools, should become a mainstream dataset.