Flame Image Segmentation Algorithm based on Background Subtraction

Considering the defect and imperfection of flame pixel point extraction and the bad environmental adaptability in the field of the present fire flame image segmentation algorithm, we put forward a kind of new algorithm based on the background difference method and fire flame color criterion. The adaptive background differencing method can detect and find objects which are moving or changing in the view field. The color criterion of fire flame can judge the color of flame of the moving objects, and then extract the flame image. Finally, the experimental results show that this algorithm has better adapt to the changing environment, and the flame extracting more accurately, perfect and stable.