Change Detection for Implanted IED Region of Interest Identification
Department
Electrical and Computer Engineering
Major
Electrical Engineering
Research Advisor
Agarwal, Sanjeev, 1971-
Advisor's Department
Electrical and Computer Engineering
Funding Source
US Army RDECOM CERDEC Night Vision and Electronic Sensors Directorate, Countermine Division, Airborne Application Branch
Abstract
Implanted Improvised Explosive Devices (IED) along the supply lines is a major source of casualties and operational delays in combat and peace keeping efforts. In this paper we develop several techniques for change detection in high-resolution mid-wave infrared (MWIR) imagery for IED detection. lEDs are often camouflaged so that it is very difficult to differentiate these from natural clutter, especially in urban environment. With the availability of images known to be clear of any lEDs, detection performance along the route can be improved significantly by looking for changes in the current imagery, with respect to the past imagery, that would not occur naturally. Several techniques have been investigated in order to improve performance in change detection which includes methods of image registration, preprocessing techniques, the method of image comparison, and assessment of high variance areas.
Biography
Thomas Woodard is a senior in Electrical Engineering at the University of Missouri - Rolla. He is an Army ROTC Cadet and upon graduation will receive his commission as a second lieutenant in the Infantry.
Research Category
Engineering
Presentation Type
Oral Presentation
Document Type
Presentation
Award
Engineering oral presentation, Second place
Presentation Date
12 Apr 2006, 10:00 am
Change Detection for Implanted IED Region of Interest Identification
Implanted Improvised Explosive Devices (IED) along the supply lines is a major source of casualties and operational delays in combat and peace keeping efforts. In this paper we develop several techniques for change detection in high-resolution mid-wave infrared (MWIR) imagery for IED detection. lEDs are often camouflaged so that it is very difficult to differentiate these from natural clutter, especially in urban environment. With the availability of images known to be clear of any lEDs, detection performance along the route can be improved significantly by looking for changes in the current imagery, with respect to the past imagery, that would not occur naturally. Several techniques have been investigated in order to improve performance in change detection which includes methods of image registration, preprocessing techniques, the method of image comparison, and assessment of high variance areas.
