- Author
- Goedeke, A. D. | Drda, B. | Healey, G. | Viglione, S. | Gross, H. G.
- Title
- Machine Vision Fire Detector System (MVFDS). Final Report. May 1990-November 1990.
- Coporate
- Donmar Limited, Newport Beach, CA
- Sponsor
- Air Force Engineering and Services Center, Tyndall AFB, FL
- Report
- ESL-TR-91-02, April 1991, 96 p.
- Distribution
- Available from National Technical Information Service
- Keywords
- fire detectors | false alarms | aircraft fires | artificial intelligence | algorithms | computers | pattern recognition
- Identifiers
- optical fire detectors; machine vision; video camera; image processing; computer classification
- Abstract
- Current UV and/or IR optical fire detectors can be deceived. A new detector based upon machine vision, image processing, video camera, pattern recognition, and microcomputer technologies provides immunity to false alarms and, at the same time, provides for the ability to detect very small fires at great distances in 1 second or less. In addition, the newly developed concept provides the fire sizes, location, and growth rate. Fires can be resolved and discriminated from bright moving lights and other phenomena. The detector knows when a fire reaches a specified threshold size and where it is located, thus allowing for the first time instantaneous dump of suppressant in selected zones where the fire threat exists. The "intelligent" machine vision fire detector system (MVFDS) can be operated as a system with multiple camera detector units, as well as with single units. It is also adaptable to work with current installed fire protection sytems. Its size is relatively smaller than current detector units and their control panels. The MVFDS concept has application to many fire, survivablity, and security problems.