Machine Vision Fire Detection Technology
Abstract
The paper will briefly outline problems with existing optical (ultraviolet and infrared) fire detectors and then presents a new type of detection technology. The new technology is called machine vision and has resulted in the development of the Machine Vision Fire Detection System (MVFDS). A prototype MVFDS will be demonstrated. It is based on work performed by the Air Force Fire Research Laboratory located at Tyndall Air Force Base. The two most important requirements of fire detectors used in ordnance operations are fast detection of a pyrotechnic/propellant fire event and reliable, false alarm-proof operation. It is of the utmost importance to identify the event in time to apply the suppressant to the developing fire event before a catastrophe situation occurs. It is also important that the detector does not false alarm to a nonfire event, which could result in an extended downtime of the fire protection system and production line, financial loss, and adverse environmental impact. It has been proven in tests with the MVFDS that there are certain unique characteristics that can distinguish fires from possible false alarm sources. This is done accurately, rapidly, and unequivocally with the algorithms/software developed in this system and implemented with the computer and camera hardware identified in this study. The MVFDS capabilities provide for a major increase in fire detector immunity to false alarms, and greater fire protection system reliability.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1994
- Accession Number
- ADA507527
Entities
People
- Robert A. Loyd