Forest Fire Detection using CNN

This paper introduces a novel fire detection system that uses Convolutional Neural Networks (CNNs) to address the inefficiencies of traditional smoke and heat sensors, which are often slow, costly, and limited in their detection capabilities. The proposed approach processes a custom dataset containing video frames from CCTV cameras to train a machine-learning model for fire detection. With extensive preprocessing to eliminate noise and irrelevant data, the system has a high detection accuracy of about 93%, hence a promising alternative for fire detection in many environments. The study highlights the need for timely fire detection, especially in scenarios such as forests where fires cause significant environmental damage and contribute to climate-related issues. The system demonstrates adaptability and scalability, capable of being implemented in homes, industries, and forests. It concludes that integrating this CNN-based framework with wireless sensors and CCTV networks could enhance detection accuracy and efficiency, potentially reducing fire-related losses and enabling quicker responses to emergencies. Keywords: Fire detection, Convolutional neural networks, Machine learning, CCTV, Object detection.