Autonomous long-range open area fire detection and reporting

2005 
Approximately 5 billion dollars in US revenue was lost in 2003 due to open area fires. In addition many lives are lost annually. Early detection of open area fires is typically performed by manned observatories, random reporting and aerial surveillance. Optical IR flame detectors have been developed previously. They typically have experienced high false alarms and low flame detection sensitivity due to interference from solar and other causes. Recently a combination of IR detectors has been used in a two or three color mode to reduce false alarms from solar, or background sources. A combination of ultra-violet C (UVC) and near infra-red (NIR) detectors has also been developed recently for flame discrimination. Relatively solar-blind basic detectors are now available but typically detect at only a few tens of meters at ~ 1 square meter fuel flame. We quantify the range and solar issues for IR and visible detectors and qualitatively define UV sensor requirements in terms of the mode of operation, collection area issues and flame signal output by combustion photochemistry. We describe innovative flame signal collection optics for multiple wavelengths using UV and IR as low false alarm detection of open area fires at long range (8-10 km/m 2 ) in daylight (or darkness). A circular array detector and UV-IR reflective and refractive devices including cylindrical or toroidal lens elements for the IR are described. The dispersion in a refractive cylindrical IR lens characterizes the fire and allows a stationary line or circle generator to locate the direction and different flame IR “colors” from a wide FOV. The line generator will produce spots along the line corresponding to the fire which can be discriminated with a linear detector. We demonstrate prototype autonomous sensors with RF digital reporting from various sites.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    8
    Citations
    NaN
    KQI
    []