Videotracking of Airborne Toxins

The identification and detection of chemical plumes is currently a very important and actively studied problem. Its significance lies in the fact that such information is handy in a variety of situations, such as overseeing chemical production plants and choosing evacuation routes during a terrorist action or an industrial accident. Studentes working on this project will work with data from the Dugway Proving Ground in Utah where tests are made using surrogate chemicals with similar signatures to dangerous chemicals. Multispectral video image data provided by the government will be studied by the students who will develop algorithms to identify the plume. It is crucial that any toxin is detected with 99% probability, as nondetection of dangerous elements can pose danger to lives. Also, the goal is to have practically no false alarms as having even a few of them can lead to inactivity in the event of an actual emergency. Last but not least, note that this problem involves time. We are not only interested in the identity and the concentrations of gases in the atmosphere but how and where they are moving. This is not at all a simple task, especially since gases tend to diffuse and thus make it even harder to detect and track them.