Fire Detection and Fire-Fighting Drone Using Deep Learning Technique

This paper proposes an ideal solution to detect and put out a fire by using a deep learning technique for fire detection. In this paper, the Fire-Fighting Drone is designed and built up to detect fires using CNN and Inception v3 models for object and fire detection techniques and automatically extinguish the fire. Firefighting balls used are plastic fire extinguisher balls that can be carried out by the Drone. Drones are becoming more and more viable for people to use. They are useful tools that can reach any firing point before spreading easily. In this paper, a CNN base model was created, which contains three conventional layers, three Pooling layers, and three Dense layers. Moreover, a dropout was added, and Adam optimizer and categorical cross-entropy loss function were used here. An experimental analysis was done through diverse types of platforms. The empirical results showed that the proposed design system was suitable to detect and fight fires spreading in any area.