Early Forest Fire Recognition using Deep Learning

Abstract

Currently many research organizations are under going to develop a reliable and an efficient methods to predict the fire disasters, which may urge to serious social impact as well as heavy property loss. From a video sequence of fire, we are going to implement deep learning method which counterfeit humans. In traditional method they employed human as inspector to detect the forest fire, but human resource is very expensive and also such approach has very low efficiency. To detect the particles like smoke or fire, temperature, relative humidity the fire sensors are implied. To achieve high accuracy rate of fire, it can be achieved by applying deep recurrence neural network. Here, the fire soldiers can move from the other side of the forest to save the forest.