Smoke detection for early fire-alarming system based on video processing

The paper presents a smoke-detection method for early fire-alarming system based on video processing. The basic strategy of smoke-pixel judgment is composed of two decision functions: a chromaticitybased decision function in static characteristic and a diffusion-based decision function in dynamic characteristic. The chromatic decision rule is deduced by grayish color of smoke and the dynamic decision rule is dependent on the diffusion attribute of smoke. To give an appropriate alarm, a fuzzy scheme is introduced to enhance the reliability of alarming by checking if the quantity of smoke extracted increases with the checking time. Experimental results show that the proposed method can provide an early alarm at a lower false alarm rate before the fire burns up, and hence is very attractive for the important places, social security, commercial applications, and so on.