Smoke Plume
As with flame, Digital Signal Processing (DSP) is central
element of smoke detection algorithms. Video images are processed by
series of patented digital so called excitation-aging™ filters that provide the distinctive patterns for
formation and evolution of smoke plumes. As shown by color-coded image
below, the plumes produced by smoke are clearly visible, and quite unique.
Here the original video frame with smoke plume highlighted by
SigniFire™ algorithm is shown in overlay image on the
left, while effects of the DSP filtering is shown on color coded image on the
right.
Growing smoke clouds processed by
excitation-aging™
filter produce very distinctive patterns that are identified by SigniFire™
sophisticated neural network pattern recognition algorithms.
Typically the ability to detect smoke is measured
by an obscuration value. This value is measured using a laser and relates to the
ratio of the laser intensity at the time of detection over the intensity of the
laser with no smoke present. The ratio can be used to calculate the amount of
smoke present at the sight of a spot type detector when a detector and
obscuration meter are placed side-by-side. The obscuration value for a specific
point does not relate well to video detection systems because of the
three-dimensional field of view. Although an obscuration value can not be
associated with intelligent video methods of smoke detection; they are
capable of detecting smoke before it reaches the ceiling of an enclosure -- much
earlier then spot type detectors. In
small compartments, this saves time and reduces damage. In large facilities, where
the smoke may not reach the ceiling due to stratification, it well may be the
only early warning option. |
