When remote sensing, GIS, and data science intersect they create the emerging field of geospatial analytics, which is being applied rapidly across many diverse industries. Last month, Harris hosted its first ENVI Analytics Symposium in Boulder, Colo., where more than 100 participants from 80 organizations and 10 countries gathered for two days to explore the burgeoning commercial geospatial analytics market.
Ret. Vice Adm. Robert Murrett, former director of the National Geospatial-Intelligence Agency and now a professor with the Maxwell School of Citizenship and Public Affairs at Syracuse University, gave a keynote address on global security and GEOINT. He pointed to the importance of GEOINT when assessing a region’s stability.
Murrett began by highlighting the refugees fleeing Syria to escape ISIS. He noted nearly 4 million people have been displaced and called it the greatest humanitarian crisis since the Rwanda conflicts 20 years ago.
“The shifts we’ve had in geography there have been pretty noteworthy and keeping track of it on a daily basis has been a challenge,” Murrett said. “NGA and others have been very responsive and helpful.”
Murrett also noted the role of GEOINT in Afghanistan.
“We cannot withdraw completely without understanding the geographic implications for the many bordering countries,” he said.
He concluded emphasizing the importance of integrated analysis, humans in the loop, and government-industry partnerships.
The Art of Analysis
Dr. Andrew Marx, a professor with the Center for Information Systems and Technology at Claremont Graduate University, followed Murrett and discussed challenges associated with human rights monitoring.
The troubled region of Darfur, for example, is bigger than the state of California and too large to monitor with a high-resolution satellite, he explained. Not only is the size of the region a tremendous challenge, but so too is the need for time-series analysis to detect change.
Working with data analytics company Palantir to monitor genocide in Darfur for the U.S. Holocaust Memorial Museum, Marx devised a method of time-stamping burning villages in Darfur using 11 Landsat path rows. This method was able to determine with 90 percent accuracy when a pixel was destroyed.
“This is the point where you cue your high-resolution sensor, cue your skilled imagery analysts and say ‘you need to look at this point. Something is happening,’” Marx said.
But once you’ve cued your high-resolution satellite, how do you extract information from that imagery?
Shay Har-Noy, senior director of geospatial big data with commercial imagery provider DigitalGlobe also presented, and shared how to extract information from imagery at scale—a process he described as converting “heavy data into Big Data that somebody can actually make sense of.”
“No one uses just imagery,” Har-Noy said, and encouraged the audience to consider what other types of data sets they could include in their outputs.
“If this community just keeps being pixel torturers then we won’t get very far,” Har-Noy cautioned. “If we embrace data coming down from disparate sources then we are in business.”
Citing that 90 percent of the world’s data has been created in the past two years alone, Har-Noy also urged the audience to opt for velocity and change rather than perfection.
Dr. John Irvine, chief scientist for data analytics with the Charles Stark Draper Laboratory, echoed Har-Noy’s sentiments, referencing a commercial transformation in the way the community should think about imagery and analysis.
“We should think in terms of broad, shallow analysis of data—extracting information on the fly,” Irvine said. “ … It may not be perfect but it will be enough to be useful.”
The Internet of Things
Fred Collins, an IBM distinguished engineer, said the company has done more with geospatial data in the last two years than in the last 20.
Collins discussed IBM’s Bluemix platform-as-a-service, which allows users to create, deploy, and manage applications in the cloud. Paired with IBM’s Internet of Things service, it allows apps to communicate with and consume data collected by connected devices and sensors.
“The cool thing about Bluemix is geospatial wasn’t an afterthought,” Collins said. “It was built into it.”
The platform includes capabilities to create geospatial analytics apps very quickly, he added, such as “device recipes.” IBM is also sharing Bluemix code on Github, which Collins described as “a new paradigm” for the company.
In a separate presentation, Dr. Alex Philp, vice president of business development at Adelos, shared his thoughts on the geospatial dimension of the Internet of Things, describing how as connected devices become sensors they are also becoming IP-addressable, network-driven, spatial objects.
According to Philp, many studies predict there will be 50 billion connected devices by 2020—an estimate he considers conservative, positing there will actually be more like 250 to 300 billion in the same timeframe.
Big Data Meets Agriculture
Three companies presented their capabilities in applying unmanned aerial vehicles (UAVs) and data analytics to the agriculture market.
“Our customers don’t want planes, they don’t want pictures,” said Jason San Souci, principal geospatial consultant for PrecisionHawk. “They want answers.”
According to San Souci, UAVs are expected to be an $82 billion industry by 2025, 80 percent of which will be attributed to civilian applications. Further, 80 percent of civilian applications are expected to be agricultural.
Dr. Yann Ameho, a guidance, navigation, and control engineer with four-year-old, France-based Delair-Tech, described how his company has had success using UAV imagery to estimate corn plant density.
Thomas Harris, director of data analytics for Agribotix, concluded the series with a discussion on agricultural intelligence. Agribotix, a startup of about 15 employees, is working to increase yields while reducing the amount of water and fertilizer needed for 20 crops in 20 countries.
Although the company offers a fixed-wing Hornet UAV and an Enduro quadcopter with a modified GoPro camera, some of its customers have employed the do-it-yourself method.
“We have a bring-your-own-drone data services platform,” Harris said. “We are seeing folks from around the world upload data on all kinds of different crops.”
Agribotix’s FarmLens platform provides field health and variable application reports. Looking to the future, Harris predicts emergence maps and the counts and locations of individual crops are the future of the market.
Photo Credit: Harris