Next: Decision Fusion
Up: Detection Algorithm
Previous: Temporal Wavelet Analysis
The fourth step of our fire detection algorithm is the spatial wavelet analysis
of moving regions containing mask pixels to capture color variations in
pixel values. In an ordinary firecolored object there will be little spatial
variations in the moving region as shown in Fig. (a). On the other
hand, there will be significant spatial variations in a fire region as shown
in Fig. (a). The spatial wavelet analysis of a rectangular frame
containing the pixels of firecolored moving regions is performed. The images
in Figs. (b) and (b) are obtained after a single stage
twodimensional wavelet transform that is implemented in a separable manner
using the same filters explained in Subsection . Absolute
values of lowhigh, highlow and highhigh wavelet subimages are added to
obtain these images. A decision parameter is defined for this step,
according to the energy of the wavelet subimages:

(4) 
where is the lowhigh subimage, is the highlow subimage,
and is the highhigh subimage of the wavelet transform, respectively,
and is the number of pixels in the firecolored moving region. If the
decision parameter of the fourth step of the algorithm, , exceeds a threshold,
then it is likely that this moving and firecolored region under investigation is a
fire region.
Both the 1D temporal wavelet analysis described in Subsection
and the 2D spatial wavelet analysis are computationally efficient schemes because
a multiplierless filter bank is used for both 1D and 2D wavelet transform
computation [9,14]. Lowpass and highpass filters have weights
and
, respectively. They can be implemented by register shifts
without performing any multiplications.
The wavelet analysis based steps of the algorithm are very important in fire and
flame detection because they distinguish ordinary motion in the video from motion
due to turbulent flames and fire.
Next: Decision Fusion
Up: Detection Algorithm
Previous: Temporal Wavelet Analysis
ugur toreyin
20051127