A fire alarm vision system based on IR image processing

This paper presents a fire alarm vision system based on flame detection algorithm of infrared image. The presented flame detection algorithm measures the luminance variation of candidate flame blocks of the IR image and it determines existence of the fire according to the amount of variation of luminance. By using the test images, the simulation of flame detection algorithm was performed and implemented fire alarm vision system was tested. As a result, the evaluated correct detection ratio was 85.7%.