Fire is a fairly common disaster in the world and is a cause of personal injury and property damage; the damage and loss of life and property due to fire is enormous. In this paper, we present a method using deep learning to detect fires. In the first step, we will detect fire candidates based on characteristics such as color, flicker frequency, and brightness of the flame. In the next step, we will use convolutional neural networks (CNNs) network and long short-term memory (LSTM) network to determines whether the flames represent a true fire or a non-fire moving object. And finally, we will evaluate the fire detection experiments for real-world applications.
Paper
Full text
Investigation of Deep Learning Method for Fire Detection from Videos
Semantic Scholar · Computer Science · 2021
Abstract
Fire is a fairly common disaster in the world and is a cause of personal injury and property damage; the damage and loss of life and property due to fire is enormous. In this paper, we present a method using deep learning to detect fires. In the first step, we will detect fire candidates based on characteristics such as color, flicker frequency, and brightness of the flame. In the next step, we will use convolutional neural networks (CNNs) network and long short-term memory (LSTM) network to determines whether the flames represent a true fire or a non-fire moving object. And finally, we will evaluate the fire detection experiments for real-world applications.