Real-Time Fire Detection through HSV and GMM Integration: A Vision-Based Alternative to Traditional Sensors

Navaneeth S, Jogimol Joseph

Abstract: This research presents a novel fire detection approach using computer vision that combines the Hue-Saturation-Value (HSV) color model with Gaussian Mixture Modeling (GMM) for real-time analysis. The system processes high-definition video feeds to identify critical fire indicators including flame patterns, dynamic movement, and luminosity changes, overcoming the limitations of conventional sensor-based methods. By integrating HSV-based color thresholding with GMM's adaptive background subtraction, the solution achieves accurate fire recognition with rapid response times. The architecture is designed for easy integration with current surveillance infrastructure, providing enhanced protection for both indoor and outdoor environments.

Keywords: Fire Detection, HSV Color Model, Gaussian Mixture Model, Computer Vision, Real-Time Monitoring, Flame Detection, Color Space Analysis, Background Subtraction, Safety Monitoring