Automatic Detection of Scud Tracks from MTI Radar Data

Period of Performance: 08/05/1993 - 01/31/1994


Phase 2 SBIR

Recipient Firm

Atlantic Aerospace Electronics Corp.
6404 Ivy Lane, Suite 300
Greenbelt, MD 20770
Principal Investigator


In Phase I, a new class of detection algorithms was successfully developed and demonstrated, using morphology-based signal processing methods, for detection of associated discrete and extended linear objects. These were respectively towers and power lines but the algorithmic techniques are general. In Phase II the techniques will be extended and applied to the automated detection of SCUD tracks in MTI radar observations and their association with buildings or structures where the missile launchers are hidden. JSTARS was used for the first time in Desert Storm and was in a position to observe appoximately 30 SCUD launches. For well understood reasons, no SCUDs were observed but at least some data were recorded and are availalbe for analysis. In Phase II, a methodical examination of these data will be undertaken without the pressures of conflict, to try to identify SCUD observations. The data will then be used to design adaptation of the algorithms developed in Phase I to detection of SCUDs. This application was selected for Phase II because it combines the critical need for automated processing for this sensor, the importance of the target and the general applicability to solution of wide area surveillance problems being pursued by the DARPA War Breaker program. Anticipated Benefits: The proposed Phase II addresses one of the major technological problems identified in the War Breaker program as essential to success of wide area surveillance against time critical targets. As such, it is anticipated that successful achievement of the goals will directly apply to War Breaker objectives. This is essentailly a military problem but has general appliability to commercial use as well.