Authors :
Nwosu Ifeoma .L.; Alagbu Ekene E.; Ezeagwu Christopher O.; Okafor Chukwunenye S.; Obiadi Ifeanyi P.
Volume/Issue :
Volume 9 - 2024, Issue 1 - January
Google Scholar :
http://tinyurl.com/3uv8h7x6
Scribd :
http://tinyurl.com/3weaehhy
DOI :
https://doi.org/10.5281/zenodo.10639122
Abstract :
In Nigeria, over 2,000 fire outbreaks reported
resulted in ₦1 trillion worth of property damages.
Likewise, a record of 31 market fires was reported in
various states in the country. These numbers point to the
dangers posed by fires and the need of a faster detection
approach. This paper presents a computer vision-based
smoke and fire detection system that is run on an edge
server. The proposed system consists of a custom
convolutional neural network (CNN) model which is
utilized to extract features in image frames to identify
fire and smoke occurrences. The k-fold cross validation
algorithm has been proved on a simplified CNN model
which has a small number of layers in order to improve
the performance of the image classification. The
experimental analysis of the model shows that the
proposed system is capable of classifying fire images
accordingly with an ROC value of over 0.67 in each
class. This model is recommended for use in deep
learning tasks that require automatic feature extraction
and object detection in image processing applications.
In Nigeria, over 2,000 fire outbreaks reported
resulted in ₦1 trillion worth of property damages.
Likewise, a record of 31 market fires was reported in
various states in the country. These numbers point to the
dangers posed by fires and the need of a faster detection
approach. This paper presents a computer vision-based
smoke and fire detection system that is run on an edge
server. The proposed system consists of a custom
convolutional neural network (CNN) model which is
utilized to extract features in image frames to identify
fire and smoke occurrences. The k-fold cross validation
algorithm has been proved on a simplified CNN model
which has a small number of layers in order to improve
the performance of the image classification. The
experimental analysis of the model shows that the
proposed system is capable of classifying fire images
accordingly with an ROC value of over 0.67 in each
class. This model is recommended for use in deep
learning tasks that require automatic feature extraction
and object detection in image processing applications.
