Towards automatic smoke detector inspection: Recognition of the smoke detectors in industrial facilities and preparation for future drone integration

arXiv cs.LG / 3/27/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper proposes an automatic smoke-detector inspection system to improve fire-safety maintenance by enabling faster, safer, and cheaper inspections without workers needing to access hard-to-reach ceiling-mounted devices.
It focuses on the smoke-detector recognition component and shows how it can be integrated into a future drone-based inspection workflow.
The study benchmarks three object-detection approaches—YOLOv11 (including an embedded-friendly variant), SSD, and transformer-based RT-DETRv2—across different backbone sizes.
Because real-world training data is difficult to collect, it evaluates training strategies that combine real and semi-synthetic data plus multiple augmentation methods.
On challenging test conditions (e.g., motion blur, small/partial detector views), the best result is reported for YOLOv11n with an average mAP@0.5 of 0.884, and the authors release code, pretrained models, and datasets.

Abstract

Fire safety consists of a complex pipeline, and it is a very important topic of concern. One of its frontal parts are the smoke detectors, which are supposed to provide an alarm prior to a massive fire appears. As they are often difficult to reach due to high ceilings or problematic locations, an automatic inspection system would be very beneficial as it could allow faster revisions, prevent workers from dangerous work in heights, and make the whole process cheaper. In this study, we present the smoke detector recognition part of the automatic inspection system, which could easily be integrated to the drone system. As part of our research, we compare two popular convolutional-based object detectors YOLOv11 and SSD widely used on embedded devices together with the state-of-the-art transformer-based RT-DETRv2 with the backbones of different sizes. Due to a complicated way of collecting a sufficient amount of data for training in the real-world environment, we also compare several training strategies using the real and semi-synthetic data together with various augmentation methods. To achieve a robust testing, all models were evaluated on two test datasets with an expected and difficult appearance of the smoke detectors including motion blur, small resolution, or not complete objects. The best performing detector is the YOLOv11n, which reaches the average mAP@0.5 score of 0.884. Our code, pretrained models and dataset are publicly available.