Can I read Fire Detection Alarm System Using Deep Learning on EtoBox?

Fire Detection Alarm System Using Deep Learning by Shivansh Uppal; Supriya Raheja; Nayan Ranjan Das is a scholarly article available to read on EtoBox.

What is Fire Detection Alarm System Using Deep Learning about?

Fire alarm systems designed with Deep Learning algorithms are more sophisticated than traditional fire alarm systems in terms of saving lives. Isolated sensors have traditionally been used to detect fires, but they are incapable of determining the extent of the fire and informing disaster preparedness teams. To address this vulnerability, this research presents an intelligent fire detection technology that not only identifies flames using photos but also alerts the user with a 4500Hz siren that buzz's for around 2500ms. For this work, the use of convolutional neural networks is applied, for evaluating visual images. A structured dataset for fire and non-fire images totaling 3730 images was utilized to condense the anticipated output. The proposed work helps to reduce false alerts. This method is incredibly trustworthy. Various callbacks, such as EarlyStopping and ReduceLROnPlateau, contributed to the accuracy, with the added benefit of lowering the likelihood of model overfitting. Convolutional Neural Network results have a better accuracy of 93.08% over other machine learning techniques namely AdaBoost and Linear Regression.

Author

Shivansh Uppal; Supriya Raheja; Nayan Ranjan Das

Publisher

IEEE

Published

2023

Language

EN