A new intelligent fire color space approach for forest fire detection

Abstract

In addition to the tragic loss of life, precious natural and personal property, forest fires pose a huge threat to ecologically healthy forests and environmental protection. A forest-fire-identification is explored in this article. Because the color space model FCS brings a high false alarm rate in flame recognition, this paper proposes an improved flame recognition color space (IFCS) based on chaos theory and k-medoids particle swarm optimization algorithm. The use of IFCS color space for flame recognition can ensure simple and fast calculations and more prominent flame/non-flame pixel color attribute difference characteristics when it is compared to FCS. The IFCS flame recognition color space is obtained by using methods such as initializing the particles in the chaotic sequence, adaptively adjusting the inertia weight, dynamic nonlinear adjustment of the learning factor, and jumping out of the local optimum from the chaotic search. In the IFCS flame color space, the binary image is obtained by the classic Otsu threshold method, and the Flame Recognition algorithm (IOFR) algorithm is established based on IFCS and Otsu. The experimental results show that based on the FCS flame recognition algorithm, the IOFR algorithm effectively reduces the flame misjudgment rate.