Multi-sensory, Data Fusion Detection System for Improved Situational Awareness

Abstract

A multi-sensory approach is being used to develop new detection capabilities for improved damage assessment and real-time situational awareness. The detection system combines surveillance camera video images with selected spectral and acoustic signatures and image recognition technologies to provide a broad range of situational awareness. Various spectral and acoustic signatures, new video imaging techniques, and image recognition methods have been investigated and integrated into a multi-sensory prototype system. The prototype system is able to detect event signatures within the volume of a space (i.e., a "volume sensor") rather than relying on spot-type fire detectors. Two prototype systems were built and assessed in full-scale testing aboard the ex-USS Shadwell side-by-side with two commercial video image fire detection systems and several spot-type fire detection systems. Tests included a wide range of fire and nuisance sources, plus flooding and pipe rupture scenarios under actual shipboard background conditions. The prototype systems are shown to outperform the commercial fire detection systems for flaming and smoldering fires with a high level of nuisance immunity. In addition, they successfully detected the pipe ruptures and flooding scenarios. The system can be adapted for homeland security.