An Agent-based Algorithm for Forest Fire Detection

For remote sensing image analysis, it is necessary to use spatial information surrounding pixels, in addition to spectral information of them. Several works have been done in this field. On the other hand, autonomous agents, a newly explored area of research in image processing, can operate directly in the two-dimensional lattice of a digital image and provide an attractive abstraction to encapsulate local and intelligent processing. Our purpose is to study the potential of agents to be applied in processing of remote sensing imagery. Therefore, in this paper, we present an agent based approach for forest fire detection. It utilizes the tests used in MODIS version 4 contextual fire detection algorithm to determine agent behavioral responses. To do so, a collection of distributed autonomous agents will operate at the individual pixels of the image divided into 10×10 pixel blocks. Performance of the proposed algorithm is compared with MODIS version 4 contextual fire detection algorithm and ground-based measurements. The results show good agreement between the algorithms and field data.