Title: Fuzzy based Manu's fire segmentation algorithm
Authors: Pratik Chhetri; P.K. Nizar Banu
Addresses: Department of Computer Science, CHRIST (Deemed To Be University), Bangalore, India ' Department of Computer Science, CHRIST (Deemed To Be University), Bangalore, India
Abstract: Forests are the primary source of natural resources and oxygen of the planet, thus any harm to the forest imbalances the natural environment. One of the primary reason for this imbalance are forest fires which requires a rigid fire surveillance system to detect fire effectively. The objective of this paper is to build a better algorithm for fire region segmentation regardless of different light illumination conditions. The algorithm uses CIE L*a*b* color space which plays a crucial role to scrutinise each fire pixel of the image and a set of fuzzy value range from 0.1 to 0.9 is used to increase the rate of accuracy of segmentation. The proposed algorithm is presented in two parts, images with extra background light source and images with low background light source which makes the segmentation process robust. A synthetic dataset created and tested with the algorithm and the results are much satisfying.
Keywords: amazon rainforest fire; CIE L*a*b*; set of fuzzy values; fuzzy logic; fire region segmentation; luminosity.
DOI: 10.1504/IJIEI.2020.111252
International Journal of Intelligent Engineering Informatics, 2020 Vol.8 No.3, pp.221 - 238
Received: 27 Apr 2020
Accepted: 08 Jul 2020
Published online: 16 Nov 2020 *