Crisis Mapping

One of the important functions of the Joint Research Centre (JRC) of the European Commission (EC) is aiding relief efforts in humanitarian crises and natural disasters around the world with accurate information based on analysis of remotely sensed data.

The Global Security and Crisis Management Unit in Ispra, Italy, includes remote-sensing experts, computer scientists, visual interpreters, crisis mappers, and political interpreters. Together they analyze satellite and aerial images from a variety of sources and combine the information with other inputs, such as social media, to develop a comprehensive view of the situation on the ground.

Their work can provide relief workers with advance information about damaged infrastructure like buildings and road networks as well as an estimate of the number of people needing help. As one of the world’s major providers of such relief, the EC depends on accurate information, says Georgios Ouzounis, a computer scientist at the JRC.

Ouzounis says that one of his challenges is to design modules capable of dealing with massive data sets in a short period of time. The variety of acquisition modes means that “our algorithms have to rely mostly on geometry and not so much on radiometry. This is a critical detail since we are asked to analyze scenes anywhere around the world, anytime, in a consistent manner.”

He says that with advancing computer technology, the world’s total land mass fills about 600 terabytes of data at half-meter resolution. A standard server box can analyze all of it in about four hours and run any query in 20 minutes.

Examples of the unit’s work in recent years include providing accurate information about damaged infrastructure after the 2010 earthquake in Haiti. They provided a map to show destroyed buildings and quantify the amount of rubble on the ground. (See images below.) In refugee situations, they are able to count tents and other housing structures, allowing them to make estimates of the number of people displaced.

  A neighborhood in Port-au-Prince, Haiti, following the earthquake.

  A rubble map of the same neighborhood generated by spatial aggregation of highly textured image regions. The latter are pinpointed by querying the Alpha-Tree structure.

This masked rubble map is a high semantic layer allowing for quantitative analysis of the extent of the damage.
(Courtesy G.K. Ouzounis, M. Pesaresi, and P. Soille, JRC.)

In the tsunami disasters in the Indian Ocean (2004) and Japan (2011), they identified destroyed road networks — of primary importance in relief efforts.

A video interview with Ouzounis is available on the SPIE Newsroom at spie.org/JRC.

Two papers presented at SPIE Defense, Security, and Sensing from the JRC reported on different aspects of the unit’s work: “Improving the automated detection of refugee/IDP dwellings using the multispectral bands of the Worldview-2 satellite” and “A new compact representation of morphological profiles: report on the first massive VHR image data processing at the JRC.”

Thomas Kemper of the JRC presented a new methodology for automated detection of refugees and other displaced persons via remote sensing.

Kemper and colleagues have tested a new technique for the automated extraction of features from satellite images based on differential morphological decomposition segmentation.

Their analysis of the number of dwellings at a camp in Darfur, Sudan, used an interactive image information tool and images from the WorldView-2 satellite, with good results obtained by using all eight spectral bands from the WorldView-2 sensor.

The DigitalGlobe satellite operates at an altitude of 770 kilometers and provides 46 cm panchromatic resolution and 1.85 meter multispectral resolution. WorldView-2 has an average revisit time of 1.1 days and is capable of collecting up to 785,000 square kilometers (303,000 square miles) of 8-band imagery per day.

Changes in U.S. military needs could mean new opportunities for photonics companies, especially in the area of IR sensing technologies.

Panelists at a government funding session at SPIE Defense, Security, and Sensing (DSS) in April said companies developing new sensing devices and technologies for improving image resolution for soldiers would find a welcome audience among U.S. defense agencies.

New and stringent requirements for modern asymmetric warfare are “good for business,” according to A. Fenner Milton of the U.S. Army Research, Development and Engineering Command. Milton, who received the 2010 SPIE DSS Lifetime Achievement Award, is a program committee member for the SPIE Infrared Technology and Applications conference at DSS.

“The nature of intelligence and surveillance, our ‘mission pull,’” added Brent Appleby of DARPA, “now involves looking for people rather than objects and for things that are well hidden.” But current sensor technology tends to be better at seeing objects rather than people, which limits effectiveness, Appleby said.

More information: optics.org/news/3/4/40

Solar panels at two fish hatcheries in Haiti are contributing to economic prosperity and food security for the impoverished country still rebuilding after a devastating 2010 earthquake.

Above, 63 PV panels and a battery system provide around-the-clock electricity for an air filtration system at the Lashto Fish Farm in Haiti.

The panels provide nearly 15 kW of electricity at the Caribbean Harvest Foundation’s fish hatchery, enabling increased production of tilapia and additional fish-farming jobs.

A solar system at another fish hatchery in Croix de Bouquets, Haiti, will have 400 solar panels when completed.

Among the organizations supporting these and other solar-energy projects in Haiti are Trina Solar, NRG Energy, the Solar Electric Light Fund, and the Clinton Global Initiative.

To receive a print copy of SPIE Professional, the SPIE member magazine, become an SPIE member.