SCIENCE & ENGINEERING NEWS

Los Alamos, N.M. — The U.S. Department of Energy’s Los Alamos National Laboratory is using a sophisticated image analysis technology to create high-resolution maps of the destruction caused by the Cerro Grande wildfire.

The computer program, called GENIE for GENetic Imagery Exploitation, is helping environmental restoration scientists pinpoint small stands of burned trees, track watershed damage and create detailed maps of new vegetation.

Members of the Nonproliferation and International Security Division’s Space and Remote Sensing Sciences group created GENIE in 1999.

“Using GENIE to analyze various kinds of image data collected during and after the Cerro Grande Fire demonstrates how a technology developed for one use can be applied to other situations. GENIE wasn’t developed for wildfire restoration applications, but it has certainly worked well,” said John Szymanski, one of GENIE’s developers and co-leader of the GENIE team.

The GENIE software system is rooted in evolutionary computing, a growing field of computer science. Evolutionary computer programs at Los Alamos mimic the evolution of biological populations in nature.

GENIE is a kind of computerized “survival of the fittest.” GENIE generates an initially random population of software algorithms, and then checks each algorithm’s “fitness” for use against the problem under consideration. The less fit algorithms are eliminated, and the more fit algorithms are allowed to mutate and “reproduce” with each other – often by combining sections of code from two relatively successful algorithms – to create a new population which goes through the fitness evaluation again, repeating the cycle (or generation) many times until a good algorithm candidate is found. That algorithm is then used to solve the problem under consideration.

To map the destruction caused by the Cerro Grande Fire GENIE used image data provided by the U.S. Forest Service, the DOE’s Remote Sensing Laboratory and the USGS’s LANDSAT satellite. The data that was provided by the Remote Sensing Lab in late May using the Daedalus multispectral scanner was particularly beneficial.

The GENIE team created burned-area maps of the terrain around Los Alamos for use by the Burned Area Emergency Rehabilitation team, a collaboration made up of representatives from the United States Forest and National Park services, the Bureau of Land Management, the United States Geological Survey and the Laboratory. BAER will use the maps to guide environmental restoration efforts around Los Alamos.

The GENIE software combines an easy-to-use, interactive browser-like screen with parallel and scalable processing. This means that almost anyone can be trained to use GENIE in just a few minutes on computers ranging in size and power from desktop systems to supercomputers.

Some areas, besides mapping, in which the GENIE application might be used include film restoration, medical image analysis, and the detection of space physics phenomena. While many of these are still in the early investigation stage, some seem to show significant promise.

The technology also has applications far beyond Earth. Using data provided by Mars Global Surveyor, GENIE counted a field of craters on Mars, successfully demonstrating that GENIE’s uses range from space-based commercial mineral prospecting to planetary cartography. Its creators envision GENIE automatically sorting through mountains of data collected from Mars and other planets to locate specific mineral resources or geographic features.

“Imagine feeding reams of data for a given planet into a supercomputer for analysis by GENIE and then going home,” said Jeff Bloch, a co-leader of the GENIE team. “You could provide GENIE with an example of what you are looking for and come back the next day and automatically have a map showing nearly all the mineral resources, canyons, craters, golf courses or whatever features.”

For more Los Alamos news releases, visit World Wide Web site http://www.lanl.gov/external/news/releases .

============================================================