Spectral-Based Volume Sensor Testbed Algorithm Development, Test Series VS2

Abstract : The Advanced Volume Sensor Project is one element of the ONR Advanced Damage Countermeasures FNC program. The Volume Sensor Project is developing new methods for remote situational awareness and damage control event detection using conventional video cameras and other techniques. The Spectral-Based Volume Sensor (SBVS) Testbed uses a suite of single-element optical detectors operating outside the visible region. Event detection algorithms were developed to make use of the data generated by the SBVS Testbed. These algorithms detect flaming and smoldering sources both within and outside the SBVS Testbed's field of view (FOV). A positive nuisance classification algorithm was also developed for arc welding and similar nuisance events, which detected all welding events with extremely few false alarms. The developed algorithms were tested against data collected during the Volume Sensor Test Series 2, July to November 2003. Comparable performance to COTS OFDs was achieved for FOV flaming sources, and superior performance was demonstrated for partially or completely obscured flaming sources.