DESIGN AND IMPLEMENTATION OF A LIGHT WEIGHT CNN MODEL FOR FIRE DETECTION

When compared to traditional fire recognition frameworks based on sensors, vision-based fire detection frameworks have recently gained popularity. The need for video perception in private, modern, business areas, and wooded areas has expanded the application of a vision-based fire system. Recently, a number of fire-related accidents have occurred as a result of inadequate surveillance or the inability to cover those uncertain regions such as restricted areas in forests or factory buildings. To avoid such mishaps, we propose a method based on Convolution neural networks.