Published Paper Details:

Fire is one of the hazardous, unwanted and sometimes uncontrolled events. Urban fires generally occur because of short-circuits in electrical devices, leakage on LPG gas cylinder ducts, or due to human ignorance itself. But at present, the system that is generally used in the community is a smoke sensor that is affixed to the ceiling of the room to detect a fireplace. Apart from the matter of early fire detection, present fire alarm systems are inefficient in terms of the false triggering of the alarm systems. Traditional Fire detection methods that is. Sensor-based and Color Based Flame detection have high false rate and low accuracy. Real-time Fire detection using Convolution Neural Network will overcome many problems in traditional fire identification. Convolutional Neural Network (CNN) is an algorithm which classifies images with a high degree of accuracy and good performance. The proposed system will be using the Convolution Neural Network, YOLO v3 (You Only Look Once Version 3) algorithm, OpenCv, Keras deep learning library.