Issue No. 03 - May/June (2000 vol. 15)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/5254.846287
This paper presents a system that combines computer vision tools, neural networks and expert fuzzy rules for the detection of forest fires in open areas. The techniques applied are based on image processing, visual-infrared images matching, memory of previous events, meteorological and geographical information, motion, size and location. The proposed system improves the reliability of infrared detection by eliminating a large number of false alarms. The paper includes real experiments carried out in forest scenarios. The paper also includes the results of the validation of the system, which took place in the scenario of the Reserve Park of "Los Alcornocales" (Alcal? de los Gazules, C?diz, Spain) during summer of 1998. The software development and implementation of the system have been carried out as a part of the DEDICS project, funded by the European Commission in the Telematics Application Program.
Pattern recognition, image processing, fuzzy logic, neural networks, tele-detection, infrared technology, and environment protection.
B. C. Arrue, J. R. de Dios and A. Ollero, "An Intelligent System for False Alarm Reduction in Infrared Forest-Fire Detection," in IEEE Intelligent Systems, vol. 15, no. , pp. 64-73, 2000.