Volume Sensor for Shipboard Damage Control
- Directorates & Divisions
- Nanoscience Institute
- Laboratory for Autonomous Systems Research
- NRL Review
- 2011 NRL Review
- 2010 NRL Review
- 2009 NRL Review
- 2008 NRL Review
- 2007 NRL Review
- 2006 NRL Review
- 2005 NRL Review
- 2004 NRL Review
- 2003 NRL Review
- 2002 NRL Review
- Future Naval Capabilities
- NRL Research Library
- Program Sponsors
- Accept the Challenge
- About NRL
- Doing Business
- Public Affairs & Media
- Field Sites
- Visitor Info
- Contact NRL
2Hughes Associates, Inc.
3Nova Research, Inc.
Introduction: Situational awareness is an important aspect of Damage Control (DC) that requires novel approaches to meet the demands of the Navy environment. Multicriteria point-sensor arrays comprised of smoke detectors and chemical sensors have demonstrated improved response times and superior nuisance source rejection compared to individual point sensors. However, these fire detection systems depend on diffusion of the particles and chemical species to the detector, which results in slow responses to smoldering fires. In addition, they have difficulty discriminating fire-like nuisance sources, such as welding, grinding steel, and cutting with a torch. Surveillance cameras in ship spaces provide a visible image that can be used to confirm the presence of fire or nuisance sources. Some fire detection companies are investigating video image detection (VID) using machine vision for fire detection in large facilities such as warehouses and tunnels. VID is a new and very attractive approach to fire detection because of the ability to see the entire compartment or volume of a space (i.e., volume sensor). The VID system provides rapid fire detection without requiring the fire effluents to travel to the detector.
Using a multisensory approach, the Naval Research Laboratory is developing a new detection capability for DC in the shipboard environment. The Advanced Volume Sensor Project is an important element of the ONR Future Naval Capabilities program, Advanced Damage Countermeasures. This program seeks to develop and demonstrate improved DC capabilities to include anticipatory DC response mechanisms to help ensure that the recoverability performance goals for the CVN21 and the DD(X) family of ships can be met with established manning levels and systems. Figure 10 shows the overall concept. Advanced Volume Sensor is a multiyear project to identify, evaluate, and adapt video image detection technologies for improved situational awareness and damage control assessment onboard Navy ships. Various spectral and acoustic signatures are being used in combination with video images and image recognition technologies for the development of a sensor system that is able to provide a broad range of situational awareness for a space. The objective of this project is to develop an affordable, real-time, remote detection system that will identify shipboard DC conditions and provide an alarm for events such as fire, explosions, pipe ruptures, and flooding level. The approach takes advantage of existing and emerging technology in the rapidly growing fields of optics, acoustics, and computer processing. This technology uses conventional surveillance cameras, which are currently being incorporated into new ship designs, and therefore will provide multiple system functions with the same hardware.
Volume Sensor Development: Three commercial video-based fire detection systems have been evaluated onboard the ex-USS Shadwell, the Naval Research Laboratory's full-scale fire research facility in Mobile, Alabama. The commercial VID systems were evaluated to assess the state of the art and to determine suitability for shipboard applications.1 Test results indicated that the VID systems using smoke alarm algorithms could provide equivalent or improved fire detection compared to point-type smoke detectors for most of the conditions evaluated. However, improvements are necessary for robust detection in the small, cluttered spaces characteristic of the Navy environment.2,3 For example, these systems cannot detect a fire obscured from the view of the camera.
Methods to enhance the performance of VID for the Navy application and expand the capabilities, such as to include detection of flooding and pipe ruptures, are being developed. The primary areas being investigated are (1) visible image video with improved machine vision algorithms, (2) augmenting visible image video with long wavelength video and spectral sensors, (3) the addition of acoustic sensors, and (4) multivariate algorithm development. The spectral-based component is focused on optical techniques in spectral regions outside the visible region, or otherwise using methods that emphasize spectral discrimination in conjunction with the visible VID techniques. Acoustic signatures are being evaluated for enhanced discrimination of DC events, particularly flooding and pipe ruptures. Advanced algorithms are being applied to the video images to identify the alarm conditions. Multivariate data analysis is being explored to integrate the sensor components and identify a variety of DC events. A novel approach to video database retrieval and archival methods was developed. All of these methods are being evaluated using DC events including fire, flooding and pipe ruptures. The successful components are being integrated into a breadboard system for further optimization of the detection methods. Development of the machine vision and multivariate data analysis methods will be a major emphasis of future work. The intention is to provide a more comprehensive overall volume sensor than would result from video detection alone.
Two distinct approaches to optical detection outside the visible are being pursued.4,5 These are long wavelength, nightvision cameras, which provide some degree of both spatial and spectral resolution or discrimination, and single or multiple element narrow spectral band detectors, which are spectrally but not spatially resolved and operate with a wide field of view at specific wavelengths ranging from the mid infrared to the ultraviolet. The primary advantages of long wavelength imaging are the higher contrast for hot objects and more effective detection of reflected flame emission compared to images obtained from cameras operating in the visible region. Our approach exploits the long wavelength response of standard charge coupled device (CCD) arrays used in many cameras (e.g., camcorders and surveillance cameras). A long pass filter transmits light with wavelengths longer than a cutoff, typically in the range 700-900 nm. This increases the contrast for fire, flame, and hot objects, and suppresses the normal video images of the space, thereby effectively providing some degree of thermal imaging. There is more emission from hot objects in this spectral region than in the visible (< 600 nm). Testing has demonstrated detection of objects heated to 400 °C or higher.
To our knowledge, this is the first use of nightvision videos for indoor fire detection. The approach is a compromise between expensive, spectrally discriminating cameras operating in the mid-infrared (IR) and inexpensive, thermally indifferent visible cameras. Nightvision is particularly useful in several respects: providing high-contrast images and therefore straightforward line-of-sight detection for flames and hot objects, where the latter would include bulkheads that have been heated by obstructed fires (Fig. 11) and identifying obstructed fire and flame based on reflected light (Fig. 12). A simple luminosity-based algorithm has been developed and used to evaluate camera/filter combinations for fire, smoke and nuisance event detection.6 The graphical user interface for this algorithm is shown in Fig. 12.
Conclusions: New capabilities for situational awareness for shipboard DC are being developed. Spectral and acoustic signatures combined with video images and machine vision provide event recognition beyond fire detection to include flooding, and pipe ruptures. A state-of-the-art database retrieval and archival method has been developed to facilitate algorithm development and optimization. Machine-vision algorithms with faster responses to fires and fewer false alarms than currently available have been demonstrated. Fire detection has been improved through the development of a low-cost method for long wavelength imaging. Long pass filters applied to the long wavelength CCD response of surveillance cameras increases the contrast to flames and hot objects and extends fire event recognition to fires outside of the field of view of the camera and behind bulkheads. Real-time fire detection with these cameras is possible by using a recently developed algorithm based on luminosity. These techniques have applications beyond shipboard use and could be beneficial in monitoring facilities for homeland security.
[Sponsored by ONR]References
1D.T. Gottuk, M.A. Harrison, J.L. Scheffey, S.L. Rose-Pehrsson, F.W. Williams, and J.P. Farley, "An Initial Evaluation of Video-Based Fire Detection Technologies," NRL Letter Report 6180/0457, December 16, 2002.
2M.A. Harrison, D.T. Gottuk, S.L. Rose-Pehrsson, J.C. Owrutsky, F.W. Williams, and J.P. Farley, "Video Image Detection (VID) Systems and Fire Detection Technologies: Preliminary Results from the Magazine Detection System Response Tests," NRL Letter Report 6180/0262, July 21, 2003.
3D.T. Gottuk, M.A. Harrison, S.L. Rose-Pehrsson, J.C. Owrutsky, J.P. Farley, and F.W. Williams, "Shipboard Evaluation of Fire Detection Technologies for Volume Sensor Development: Preliminary Results," NRL Letter Report 6180/0282, August 25, 2003.
4J.C. Owrutsky, D.A. Steinhurst, H.H. Nelson, and F.W. Williams, "Spectral Based Volume Sensor Component," NRL/MR/6110--03-8694, July 30, 2003.
5D.A. Steinhurst, J.C. Owrutsky, S.L. Rose-Pehrsson, D.T. Gottuk, F.W. Williams, and J.P. Farley, "Spectral-Based Volume Sensor Testbed VS1 Test Series Results ex-USS Shadwell, April 20-25, 2003," NRL Letter Report 6110/075, June 27, 2003.
6D.A. Steinhurst, C.P. Minor, J.C. Owrutsky, S.L. Rose-Pehrsson, D.T. Gottuk, F.W. Williams, and J.P. Farley, "Long Wavelength Video-Based Event Detection, Preliminary Results from the CVNX and VS1 Test Series, ex-USS Shadwell, April 7-25, 2003," NRL/MR/6110--03-8733,December 31, 2003.