Beyond RGB Fire Detection: Infrared Imaging, Dataset Bias, and XAI

Dear Colleagues,

Recent advances in deep learning have significantly improved vision-based fire and flame detection in environments ranging from industrial and civil infrastructures to forests and open landscapes. Fires occurring in these settings differ substantially in terms of temperature, combustion dynamics, and electromagnetic spectral signatures, making the development of customized detection systems and processing pipelines essential. Even within similar environments and identical combustible materials, flame characteristics may vary considerably depending on the combustion stoichiometry and airflow conditions.

Despite this diversity, most existing approaches predominantly rely on RGB imagery and opaque neural network models, raising critical concerns regarding robustness, bias, and trustworthiness in safety-critical applications. In particular, the limited representation of non-canonical flames (such as blue flames) in publicly available datasets, together with the absence of systematic interpretability frameworks, significantly hinders reliable deployment in industrial and domestic scenarios.

This Special Issue aims to bring together high-quality contributions addressing multimodal fire detection using RGB, infrared, and thermal imaging, as well as explainable artificial intelligence (XAI), dataset design, and experimental validation under both controlled and real-world conditions. By combining methodological surveys, experimental studies, dataset papers, and applied case reports, the Special Issue seeks to advance transparent, reliable, and generalizable AI-based fire detection systems.

Submissions proposing novel datasets or original processing pipelines for existing datasets are particularly encouraged. Ideally, contributions to this Special Issue should include a thorough discussion of model explainability and interpretability, especially in relation to safety, bias, and failure modes in fire and flame detection tasks.

Dr. Bogdan Diaconu
Dr. Lucica Anghelescu
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 250 words) can be sent to the Editorial Office for assessment.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Fire is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• fire/flame detection
• vision-based fire recognition
• deep learning
• multimodal sensing
• infrared/thermal imaging
• RGB–IR data fusion
• explainable artificial intelligence (XAI)
• model interpretability
• safety-critical systems
• dataset bias
• blue flame detection
• robustness and generalization
• experimental validation
• literature measurement and analysis of research trends