engleski

Extraction of information about spatial characteristics of fire plume from a video signal

This work presents an instrumentation system developed for an on-line dimension measurement of fire plume from a color video signal. The dimensions of a flame (high, width and sum-total of area) are important parameters of a combustion process of burning solid fuels and an input parameter for the calculation of flame flickering (oscillation frequency of the flame). The systems consists of a CCD camera or tape video recorder that records information about fire and a PC computer with a frame grabber and associated image processing software. The method for extraction area of fire plume from background in a flame image (frame) is based on different methods of filtering. The first method is luminance filtering (setting the threshold and Wavelet decomposition of an image). The second method is hue filtering by setting a threshold for the hue component of an image data. On condition that the appropriate thresholds for the filtering process could be defined, the separation of the flame and background can be accurately performed. However, it is impossible to make a perfect separation using only a video signal, even though a noise reduction allows us to calculate a flame dimensions with an appropriate precision. The developed system has been utilized to investigate the relationship between of three distinct regimes of the fire plume (persistent, intermittent and buoyant zones of the natural diffusion flame).

spatial characteristics of fire plume; video signal; Wavelet decomposition of an image; Daubechies and Haar wavelet family; luminance and hue filtering of an image.

nije evidentirano