Applications of Computational Verb Theory to the Design of Accurate Video Flame-detecting Systems

The conventional video flame-detecting systemshave very high false alarm rate, low ability of suppressinginterferences and cannot detect controllable flames. Therefore,these systems cannot be used in open spaces, forests and otherplaces where light inferences exist. Also, these systems cannotprevent flame at the very earlier stage when the flames arecontrollable. The newly designed video flame-detecting system isbased on computational verb theory. By using physical linguisticnetworks, the cognitive models of flames are built to embedhuman perceptions of flames in natural languages into computeralgorithms. This comprehensive model encoded the textures,structures, spatial-temporal dynamics of flames into a physicallinguistic network. By doing so, the image understanding problemof flames was transformed into a physical linguistic model whichhas high efficiency, big flexibility, and good generalities. Inthis paper, the design of video flame-detecting systems fromengineering point of view will be surveyed in details.