ENRICHING FIRE DETECTION THROUGH MULTI- CLASSIFIER AND DEEP LEARNING MODEL


Saiprasad Bhise, Rachika Waghmare, Yash Deshmukh, Shrutika Jagdhane, Prof. S. S. Raskar, Prof. J. S. Mane
Student, Savitribai Phule Pune University
Abstract
Fire is one of the most destructive forces that have been a double edged sword, while it is very useful and generates energy when controlled in an effective manner, it can be quite deadly if left unchecked. The fire is combustion and a conversion and release of energy which is violent and has the ability to unleash massive amounts of destruction. This is highly undesirable circumstance that could lead to a potentially catastrophic occurrence. It has been evident in the recent years with the large number of devastating wildfires that caused a large scale loss of life and property, a large number of species of flora and fauna were extinct which one of the most deadly occurrences. The main problem is the lack of an effective and useful fire detection approach. Therefore, this research article defines an effective multi-classifier approach for fire detection that identifies the color of the fire, shape of the fire and the movement of the fire, along with the detection of smoke by using the convolution neural network. This approach has been one of the most effective techniques for the fire detection which is evident through the extensive experimental results that signify the superiority of the proposed multi-classifier fire and smoke detection approach.
Keywords: Convolution neural network, Multi-classifier, Image Morphology, Temporal effect
Journal Name :
EPRA International Journal of Research & Development (IJRD)

VIEW PDF
Published on : 2022-05-02

Vol : 7
Issue : 4
Month : April
Year : 2022
Copyright © 2024 EPRA JOURNALS. All rights reserved
Developed by Peace Soft