Vision-based smoke detector

Previous studies have documented the significant applications of the electronic smoke detector. With the capabilities of vision based fire detection and increase in the number of surveillance cameras, a lesser attention is given to the vision-based type smoke detector. Moreover, some drawbacks have...

Full description

Bibliographic Details
Main Authors: Abdullah, Ali Mohammed Noman, Htike@Muhammad Yusof, Zaw Zaw
Format: Article
Language:English
English
Published: Science Publishing Corporation 2019
Subjects:
Online Access:http://irep.iium.edu.my/73168/
http://irep.iium.edu.my/73168/1/Vision%20Based%20Smoke%20detector%20%282%29.pdf
http://irep.iium.edu.my/73168/2/Vision-based%20acceptance.pdf
Description
Summary:Previous studies have documented the significant applications of the electronic smoke detector. With the capabilities of vision based fire detection and increase in the number of surveillance cameras, a lesser attention is given to the vision-based type smoke detector. Moreover, some drawbacks have been identified in the accuracy and efficiency of smoke detection. The present study proposes a vision based smoke detector to overcome the shortcomings of the current traditional electronic and vision based smoke detectors. A Convolutional Neural Network is used to classify the smoke regions. After testing the proposed method, the accuracy was approximately 94%. When a modern approach of object detection is used to support image classifying, its accuracy increases by 96%.