CNN Based Fire Detection System

There has been a tremendous amount of work which has been done in the field of Fire detection Technology. Talking about the traditional fire detectors they have many faults such as the response time is very high which will delay the evacuation process and leads to lot of destruction and loss of huma...

Full description

Saved in:
Bibliographic Details
Published in2022 4th International Conference on Advances in Computing, Communication Control and Networking (ICAC3N) pp. 980 - 985
Main Authors Pal, Rishav, Singh Rawat, Vishal, Kumar Dubey, Vivek, Chauhan, Mukesh
Format Conference Proceeding
LanguageEnglish
Published IEEE 16.12.2022
Subjects
Analytical models
ANN
CNN
Costs
Data models
Delays
Detectors
Inception V3
RNN
Training
Transfer learning
YOLO
Online AccessGet full text

Cover

More Information
Summary:There has been a tremendous amount of work which has been done in the field of Fire detection Technology. Talking about the traditional fire detectors they have many faults such as the response time is very high which will delay the evacuation process and leads to lot of destruction and loss of human lives also these fire detector systems can not detect the smouldering fire and also the cost of installation is very high. So as to mitigate these we have come with the CNN based fire detection which solves all the problems as this system has less response time as compared with other fire detector and also the installation process and cost is less. We have made a CNN based fire detector which uses Inception V3 as a transfer learning module. The reason for inducting Inception V3 instead of AlexNet, GoogleNet, and YOLO is that it can handle large data and also can detect true and false fire and also the accuracy is good and loss is less.
DOI:10.1109/ICAC3N56670.2022.10074585