Fire detection and recognition technologies for fire safety engineering: A bibliometric and scenario-based review

Early and reliable fire detection is essential for suppression, evacuation, smoke control, and emergency response, yet research has become fragmented across computer vision, remote sensing, unmanned aerial vehicle monitoring, sensor networks, and edge intelligence. This review combines bibliometric mapping with scenario-based synthesis to examine 2450 Web of Science Core Collection records published from 2000 to 2025. CiteSpace was used to analyze publication trends, collaboration networks, keyword evolution, and burst terms, while representative studies were evaluated from a fire safety engineering perspective. Four development stages and five major application pathways were identified: image/video flame and smoke detection, satellite wildfire monitoring, unmanned aerial vehicle and airborne detection, indoor and special-scenario sensing, and datasets and evaluation systems. The field is shifting from isolated target recognition toward scenario-adaptive fire safety perception. Future work should prioritize alarm latency, false-alarm resistance, missed-detection risk, cross-scenario robustness, edge deployment, and integration with fire protection and emergency-response systems.