BOZEMAN, Montana ­ - Visual Learning Systems (VLS) of Missoula, Mont. has created a software product called Feature AnalystTM that improves the efficiency of extracting geospatial information from high-resolution satellite imagery by a factor of 50 to 200 over current methods. This technology is so noteworthy that NASA has awarded VLS over $700,000 since September to pursue the commercial application of the software for NASA's Earth Science Enterprise program, and the Department of Defense has awarded the company $100,000 for the same purpose.

Feature Analyst combines a sophisticated "adaptive interface" with machine learning technology (so-called "smart" technology). The result is a product that can identify and even distinguish fine details of objects located on Earth that have been captured by satellites located 150 to 500 miles above the surface of the earth. This technology automates a task once performed with limited accuracy and at much slower rates of speed, by humans. Feature Analyst technology is highly prized in the geographic information systems (GIS) industry, the military, and other industries that require rapid, accurate identification and location of geographic features. For example, Feature Analyst can quickly interpret satellite data to identify and distinguish between features such as roads, buildings, hydrology, and vegetation. The speed and accuracy with which this software operates makes VLS a leader in its field, but the smart technology will make Feature Analyst a leading software brand. The company plans to release the software to the public on November 1, 2001.

A unique feature of Feature Analyst is its ability to "learn" from its mistakes. Satellite images are collections of raw data that are made sensible by identification of recognizable patterns or characteristics. From an aerial photograph, for example, the human eye can easily identify a road because the eye recognizes certain characteristics, such as lines and color that enable us to recognize the feature as a road. The software functions much the same way, analyzing raw data and searching it for groups of recognizable characteristics that form a pattern that can be identified as a specific geographic feature, such as a hill or building. Like the human eye, the software can be "fooled" by the interplay of light or form, and like the human eye, Feature Analyst can be "trained" to recognize its mistakes and learn from them, which is what gives this product a leg up on its competition. Each time the technology learns from its mistakes, it gets smarter, which translates as speedier and more accurate. "The Feature Analyst is designed to model the feature extraction process by learning from examples provided by the human analyst," said CEO and Chief Technology Officer Dr. David Opitz. "The adaptive nature of our proprietary learning algorithms allows the software to improve its performance as new information is presented by the user."

VLS will use its awards from NASA and the Department of Defense to develop Feature Analyst as a plug-and-play module for the market leading GIS software. In 1999, VLS received a Phase I NASA Small Business Innovation Research (SBIR) award to develop the prototype of Feature Analyst. The most recent award from NASA is a Phase II SBIR award, which will help the company develop Feature Analyst for NASA's Earth Science Enterprise (ESE) program, which is NASA's Earth exploration initiative.

NASA has been collecting and interpreting data about Earth from the perspective of space since the first aerospace vehicle was launched. Such data supports scientific research that forms a basis of support for the development of global environmental policy. VLS's Feature Analyst is a perfect example of a software application that can help increase the efficiency of interpretation of some of NASA's vast collection of raw data.

NASA is establishing partnerships with industry, state and local governments, and other Federal agencies to pursue a broader range of applications for its Earth Science Enterprise program. For example, some uses of ESE data include developing maps from satellite images that show the condition of crops and vegetation, the potential for forest fire outbreaks, predict weather events, and measure sea surface temperature and currents. Its long-range mission is to develop an understanding of the total Earth system and the effects of natural and human-induced changes on the global environment.

VLS's Feature Analyst software supports initiatives such as NASA's ESE program and, with the projected high growth in the GIS industry, is currently positioned to capture a large share of the commercial software market. Feature Analyst is a powerful yet easy-to-use software application. The company plans to combine its traditional software development business with an online business-to-business web service to build VLS's brand presence and drive Internet community participation. VLS is based in Missoula, Mont. and is currently working with the Missoula Area Economic Development Corporation (MADEC), which is developing an incubator facility for early stage companies through a partnership with the University of Montana. According to Dick King, MADEC's Executive Director, "VLS is a great example of the type of company we hope to assist."

The TechLink Center located at Montana State University in Bozeman, linked VLS with NASA and the Department of Defense and is currently supporting VLS's application for a research agreement with NASA and its application for additional SBIR and other awards. NASA, the Department of Defense, and other federal agencies fund TechLink to assist companies in the northwestern United States with accessing, developing, and commercializing technology in partnership with federal research labs. Additional information on Feature Analyst software and Visual Learning Systems, Inc. can be obtained at www.FeatureAnalyst.com

Contact: 
Will Swearingen 
MSU TechLink Center
(406) 994-7704
wds@montana.edu

Stuart Blundell
VLS
(406) 443-0766
sblundell@intgeo.com

Back to Top