International Research Journal of Education and Technology

The primary goal of this paper is the early detection of fire in the forest region. It detects the fire by extracting   the images and videos in the camera. The research is primarily concerned with image recognition methods. This involves a series of steps, starting with the background subtraction because fire boundaries continuously changes. And then color segmentation model is used to mark the candidate regions. Finally, Convolutional Neural Network(CNN) is used for classifying the candidate regions has actual fire or non fire. Whenever actual fire incident happens it automatically detects the fire without any sensor device and sends sound alert to the forest department .

Keyword :Forest fire detection, Camera analysis, Image recognition ,Convolutional Neural Network, Machine learning.