Artificial Intelligence for Building Fire Detection and Prevention: Research Progress, Applications and Future Trends (2016–2026)

Abstract

Traditional fire detection technologies for buildings are no longer adequate for the fire prevention and control needs of complex structures, while emerging artificial intelligence technologies have become the core path to break through the bottlenecks in this industry. Existing reviews suffer from non-standardized paradigms, one-sided scopes, insufficient methodological evaluation, incomplete time coverage, and a lack of engineering orientation, failing to meet the evidence-based research needs of the field. This study strictly followed the PRISMA 2020 systematic review guidelines, selected relevant SCI papers from the Web of Science Core Collections from 1 January 2016 to 30 March 2026, in JCR Q1 and Q2, and finally included 221 valid papers; it systematically carried out bibliometric analysis and technical system sorting. The results showed that (1) the number of publications in this field showed a significant exponential upward trend, with China accounting for 52% of the research output, ranking first in the world, and (2) convolutional neural networks and YOLO series algorithms are the mainstream application technologies in the field. The study clarified the performance differences, advantages, and disadvantages, and applicable scenarios of various artificial intelligence algorithms. This study identified the existing technical and methodological limitations in the field and explored the core research directions for the future, and provided evidence-based support for the academic research and engineering implementation of artificial intelligence in the field of building fire detection and prevention.