Advanced CCTV Analytic Solution for Fire Detection

Computer Science Engineering

In order to reduce the hazards that fire accidents bring to human life, property, and the environment, effective fire detection systems are essential. Closed-circuit television (CCTV) cameras, which are common in many situations, are a useful tool for improving fire detection. This study performs a thorough investigation of cutting-edge methods for utilizing CCTV footage for enhanced fire detection, concentrating on CNN and CNN Inception architectures. The comparative performance analysis of these two convolutional neural network (CNN) models is the focus of the study. Key measures like accuracy, precision, recall, and F1 score are used for evaluation after being trained on a large dataset. The outcomes highlight the benefits of the CNN Inception model. The CNN model's accuracy of 0.6802 is surpassed by its accuracy of 0.815. Impressively, the CNN Inception model surpasses the CNN model's 0.809 precision mark with a precision of 0.909. The CNN model, on the other hand, achieves a recall of 0.7, somewhat better than the CNN Inception model. As a result, the CNN Inception model's F1 score is 0.79, while the CNN model's F1 score is 0.70. In conclusion, the article offers a thorough comparison of CNN and CNN Inception models for fire detection via CCTV footage analysis, providing relevant details about each model's individual capabilities. These results support the careful choice and enhancement of deep learning models for reliable fire detection systems. The research advances the state-of-the-art in fire detection, promoting safer settings and reducing fire-related deaths

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.