Advanced modeling for fire risk evaluation and intelligent preventive system

Sabina Umirzakova; Shakhzod Javliev; Adkhambek Madaminov; Omadjon Urishev; Dilshoda Kurbonalieva; Ayhan Istanbullu; Akmalbek Abdusalomov

doi:10.65112/tcmis.10043

Authors

Sabina Umirzakova Department of Computer Engineering, Gachon University Sujeong-Gu, Seongnam-si 13120, Gyeonggi-Do, Korea https://orcid.org/0000-0002-7694-1806
Shakhzod Javliev Department of Computer Systems, Tashkent University of Information Technologies Named after Muhammad Al-Khwarizmi, Tashkent 100200, Uzbekistan https://orcid.org/0000-0003-1411-8192
Adkhambek Madaminov Department of Computer Systems, Tashkent University of Information Technologies Named after Muhammad Al-Khwarizmi, Tashkent 100200, Uzbekistan https://orcid.org/0009-0003-0716-637X
Omadjon Urishev Department of Electronics and Instrumentation, Fergana State Technical University, Fergana, 150100, Uzbekistan https://orcid.org/0009-0005-5728-554X
Dilshoda Kurbonalieva Department of Cryptology, Tashkent University of Information Technologies Named after Muhammad Al-Khwarizmi, Tashkent 100200, Uzbekistan https://orcid.org/0009-0003-3713-7422
Ayhan Istanbullu Department of Computer Engineering, Faculty of Engineering, Balikesir University, Balikesir 10145, Turkiye https://orcid.org/0000-0002-7066-4238
Akmalbek Abdusalomov Department of Artificial Intelligence, Tashkent University of Information Technologies Named after Muhammad Al-Khwarizmi, Tashkent 100200, Uzbekistan https://orcid.org/0000-0001-5923-8695

DOI:

https://doi.org/10.65112/tcmis.10043

Keywords:

Maritime fire detection, smoke recognition, edge AI deployment, fire hazard prevention, early emergency response

Abstract

Shipboard fires are among the most catastrophic maritime incidents, as dense smoke, fluctuating lighting, and camera movement often obstruct timely detection. To address these operational challenges, we propose a lightweight yet robust vision-based fire and smoke detection system tailored for vessels with limited onboard computing resources. The proposed model integrates a partially fine-tuned ResNet-34 feature extractor with a capsule-based detection head, which preserves spatial relationships critical for identifying irregular flame and smoke patterns. Directional attention mechanisms are employed to highlight subtle, low-contrast smoke formations, which frequently occur under maritime conditions such as haze or sea glare. When evaluated on a newly created marine dataset representing a variety of real-world scenarios, the model demonstrates exceptional accuracy and real-time inference performance on edge-class devices. Our solution, which incorporates the fine-tuned ResNet-34 backbone, a PAA module, and a capsule-based detection head, outperforms all comparative methods, achieving a maximum mAP@0.5 of 90.17%, an IoU of 72.58%, Precision of 92.2%, and Recall of 88.5%. These results indicate that the model excels at both precise localization and detecting subtle or partially obscured fire and smoke regions—an ongoing challenge in shipboard imaging.

Downloads

Download data is not yet available.

References

[1] P. Zhang, Z. Song, C. Li, Y. Liu, Y. Zou, Y. Zhang, et al., “A study of engine room smoke detection based on proactive machine vision model for intelligent ship,” Expert Syst. Appl., vol. 241, p. 122689, 2024. https://doi.org/10.1016/j.eswa.2023.122689 DOI: https://doi.org/10.1016/j.eswa.2023.122689

[2] D. Kim and W. Ruy, “CNN-based fire detection method on autonomous ships using composite channels composed of RGB and IR data,” Int. J. Nav. Archit. Ocean Eng., vol. 14, p. 100489, 2022. https://doi.org/10.1016/j.ijnaoe.2022.100489 DOI: https://doi.org/10.1016/j.ijnaoe.2022.100489

[3] S. Ricci, B. S. S. K. Ravikumar, and L. Rizzetto, “Fire Management on Container Ships. New Strategies and Technologies,” TransNav, vol. 17, no. 2, pp. 415–421, 2023. https://doi.org/10.12716/1001.17.02.19 DOI: https://doi.org/10.12716/1001.17.02.19

[4] A. Abdusalomov, S. Umirzakova, A. Kutlimuratov, D. Mirzaev, A. Dauletov, T. Botirov, et al., “Lightweight UAV-Based System for Early Fire-Risk Identification in Wild Forests,” Fire (Basel), vol. 8, no. 8, p. 288, 2025.

[5] S. M. Lee and J. H. Nam, “Performance-based fire and evacuation analysis for real-time response to shipboard fire incidents,” Ocean Eng., vol. 318, p. 120166, 2025. https://doi.org/10.1016/j.oceaneng.2024.120166 DOI: https://doi.org/10.1016/j.oceaneng.2024.120166

[6] J. Zhu, J. Zhang, Y. Wang, Y. Ge, Z. Zhang, and S. Zhang, “Fire detection in ship engine rooms based on deep learning,” Sensors (Basel), vol. 23, no. 14, p. 6552, Jul. 20 2023. https://doi.org/10.3390/s23146552 DOI: https://doi.org/10.3390/s23146552

[7] R. Tao, H. Ren, and Y. Zhou, “A ship firefighting training simulator with physics-based smoke,” J. Mar. Sci. Eng., vol. 10, no. 8, p. 1140, 2022. https://doi.org/10.3390/jmse10081140 DOI: https://doi.org/10.3390/jmse10081140

[8] F. Makhmudov, S. Umirzakova, A. Kutlimuratov, A. Abdusalomov, and Y.-I. Cho, “Advanced Object Detection for Maritime Fire Safety,” Fire (Basel), vol. 7, no. 12, p. 430, 2024. DOI: https://doi.org/10.3390/fire7120430

[9] H. Lee, J. H. Chung, H. Chung, J. H. Kim, Y. Yoo, G. H. Lim, et al., “AI-enhanced fire detection and suppression system for autonomous ships,” Int. J. Nav. Archit. Ocean Eng., vol. 16, p. 100628, 2024. https://doi.org/10.1016/j.ijnaoe.2024.100628 DOI: https://doi.org/10.1016/j.ijnaoe.2024.100628

[10] H. Kong and J. Lee, “Development of Sprinkler Head Integrated with Smoke Detector,” Journal of the Convergence on Culture Technology, vol. 11, no. 3, pp. 571–580, 2025.

[11] Sangermano, V., Ciaburri, M., Russo, R. and Esposito, D.P., 2024. Human factors implication in innovative strategies for containership fires prevention and management. Safety Management and Human Factors, 151(151). https://doi.org/10.54941/ahfe1005313 DOI: https://doi.org/10.54941/ahfe1005313

[12] T. N. Dinh and H. N. Van, 2025, March. Research and Development of an Early Fire Warning System for Ships Using Artificial Neural Networks. In 2025 International Russian Smart Industry Conference (SmartIndustryCon) (pp. 555-560). IEEE. https://doi.org/10.1109/SmartIndustryCon65166.2025.10986067 DOI: https://doi.org/10.1109/SmartIndustryCon65166.2025.10986067

[13] B. Porterie, N. Dizet, Y. Pizzo, J. C. Loraud, P. Boulet, A. Collin, et al., “A simulation tool to quantify the consequences of fires on board Ro-Ro ships,” Fire Technol., vol. 60, no. 1, pp. 459–499, 2024. 10.1007/s10694-023-01515-3 DOI: https://doi.org/10.1007/s10694-023-01515-3

[14] S. Sabour, N. Frosst, and G. E. Hinton, “Dynamic routing between capsules,” Adv. Neural Inf. Process. Syst., vol. 30, p. 30, 2017.

[15] J. Del-Pozo-Velázquez, J. M. Aguiar-Pérez, P. Chamorro-Posada, M. Á. Pérez-Juárez, X. Wang, and P. Casaseca-de-la-Higuera, “Smoke Detection in Images through Fractal Dimension-based Binary Classification,” Digit. Signal Process., vol. 166, p. 105346, 2025. https://doi.org/10.1016/j.dsp.2025.105346 DOI: https://doi.org/10.1016/j.dsp.2025.105346

[16] S. Kanagamalliga and S. Rajalingam, “Innovative fire and gas recognition system featuring remote monitoring and automated alerts,” Procedia Comput. Sci., vol. 252, pp. 7–14, 2025. https://doi.org/10.1016/j.procs.2024.12.002 DOI: https://doi.org/10.1016/j.procs.2024.12.002

[17] A. Mamadmurodov, S. Umirzakova, M. Rakhimov, A. Kutlimuratov, Z. Temirov, R. Nasimov, et al., “Im Cho, Y. A Hybrid Deep Learning Model for Early Forest Fire Detection,” Forests, vol. 16, no. 5, p. 863, 2025. DOI: https://doi.org/10.3390/f16050863

[18] A. Lopez-Alanis, H. De-la-Torre-Gutierrez, A. Hernández-Aguirre, and M. T. Orvañanos-Guerrero, “Fuzzy rule-based combination model for the fire pixel segmentation,” IEEE Access, vol. 13, pp. 52478–52496, 2025. https://doi.org/10.1109/ACCESS.2025.3554140 DOI: https://doi.org/10.1109/ACCESS.2025.3554140

[19] R. A. Khan, U. I. Bajwa, R. H. Raza, and M. W. Anwar, “Fire and smoke detection using two-stream spatiotemporal network,” Neural Comput. Appl., vol. 37, no. 24, pp. 1–25, 2025. https://doi.org/10.1007/s00521-025-11412-1 DOI: https://doi.org/10.1007/s00521-025-11412-1

[20] S. S. Mohamed, M. H. Mir, and T. A. Mir, 2025, January. Enhancing Building Safety Through Machine Learning-Based Smoke Detection. In 2025 Fifth International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies (ICAECT) (pp. 1-6). IEEE. https://doi.org/10.1109/ICAECT63952.2025.10958840 DOI: https://doi.org/10.1109/ICAECT63952.2025.10958840

[21] Q. X. Zhang, G. H. Lin, Y. M. Zhang, G. Xu, and J. J. Wang, “Wildland forest fire smoke detection based on faster R-CNN using synthetic smoke images,” Procedia Eng., vol. 211, pp. 441–446, 2018. https://doi.org/10.1016/j.proeng.2017.12.034 DOI: https://doi.org/10.1016/j.proeng.2017.12.034

[22] E. Casas, L. Ramos, E. Bendek, and F. Rivas-Echeverría, “Assessing the effectiveness of YOLO architectures for smoke and wildfire detection,” IEEE Access, vol. 11, pp. 96554–96583, 2023. https://doi.org/10.1109/ACCESS.2023.3312217 DOI: https://doi.org/10.1109/ACCESS.2023.3312217

[23] X. Han, N. Pu, Z. Feng, Y. Bei, Q. Zhang, L. Cheng, et al., 2024, October. Benchmarking multi-scene fire and smoke detection. In Chinese Conference on Pattern Recognition and Computer Vision (PRCV) (pp. 203-218). Singapore: Springer Nature Singapore. DOI: https://doi.org/10.1007/978-981-97-8795-1_14

[24] L. He, Y. Zhou, L. Liu, Y. Zhang, and J. Ma, “Research and application of deep learning object detection methods for forest fire smoke recognition,” Sci. Rep., vol. 15, no. 1, p. 16328, May 10 2025. https://doi.org/10.1038/s41598-025-98086-w DOI: https://doi.org/10.1038/s41598-025-98086-w

[25] N. C. Pistol, B. Alexandrescu, R. C. Constantinescu, and D. I. Săcăleanu, Integrated and Automated Fire Detection System: A Comprehensive Review of Current Challenges and Methods, Future Trends and Prospects for Emerging Electronics Engineering with AI Integration. 2024 Advanced Topics on Measurement and Simulation. ATOMS, 2024, pp. 204–207. DOI: https://doi.org/10.1109/ATOMS60779.2024.10921597

[26] Y. Sun and J. Feng, “Fire and smoke precise detection method based on the attention mechanism and anchor-free mechanism,” Complex Intell. Syst., vol. 9, no. 5, pp. 5185–5198, 2023. https://doi.org/10.1007/s40747-023-00999-4 DOI: https://doi.org/10.1007/s40747-023-00999-4

[27] X. Fan, F. Lei, and K. Yang, “Real-Time Detection of Smoke and Fire in the Wild Using Unmanned Aerial Vehicle Remote Sensing Imagery,” Forests, vol. 16, no. 2, p. 201, 2025. https://doi.org/10.3390/f16020201 DOI: https://doi.org/10.3390/f16020201

[28] B. Wang, G. Huang, H. Li, X. Chen, L. Zhang, and X. Gao, “Hybrid CBAM-EfficientNetV2 fire image recognition method with label smoothing in detecting tiny targets,” Machine Intelligence Research, vol. 21, no. 6, pp. 1145–1161, 2024. https://doi.org/10.1007/s11633-023-1445-5 DOI: https://doi.org/10.1007/s11633-023-1445-5

[29] S. Wang, Y. Zhang, T. H. Hsieh, W. Liu, F. Yin, and B. Liu, “Fire situation detection method for unmanned fire-fighting vessel based on coordinate attention structure-based deep learning network,” Ocean Eng., vol. 266, p. 113208, 2022. https://doi.org/10.1016/j.oceaneng.2022.113208 DOI: https://doi.org/10.1016/j.oceaneng.2022.113208

[30] M. Yang, S. Qian, and X. Wu, “Real‐time fire and smoke detection with transfer learning based on cloud‐edge collaborative architecture,” IET Image Process., vol. 18, no. 12, pp. 3716–3728, 2024. https://doi.org/10.1049/ipr2.13187 DOI: https://doi.org/10.1049/ipr2.13187

[31] L. Zhang, C. Lu, H. Xu, A. Chen, L. Li, and G. Zhou, “MMFNet: Forest fire smoke detection using multiscale convergence coordinated pyramid network with mixed attention and fast-robust NMS,” IEEE Internet Things J., vol. 10, no. 20, pp. 18168–18180, 2023. https://doi.org/10.1109/JIOT.2023.3277511 DOI: https://doi.org/10.1109/JIOT.2023.3277511

[32] A. Amjad, A. M. Huroon, H. T. Chang, and L. C. Tai, “Dynamic fire and smoke detection module with enhanced feature integration and attention mechanisms,” Pattern Anal. Appl., vol. 28, no. 2, pp. 1–17, 2025. https://doi.org/10.1007/s10044-025-01461-6 DOI: https://doi.org/10.1007/s10044-025-01461-6

[33] M. Shakhnoza, U. Sabina, M. Sevara, and Y. I. Cho, “Novel video surveillance-based fire and smoke classification using attentional feature map in capsule networks,” Sensors (Basel), vol. 22, no. 1, p. 98, Dec. 24 2021. https://doi.org/10.3390/s22010098 DOI: https://doi.org/10.3390/s22010098

[34] X. Li, J. Liu, Z. Tang, B. Han, and Z. Wu, “MEDMCN: A novel multi-modal EfficientDet with multi-scale CapsNet for object detection,” J. Supercomput., vol. 80, no. 9, pp. 12863–12890, 2024. https://doi.org/10.1007/s11227-024-05932-1 DOI: https://doi.org/10.1007/s11227-024-05932-1

[35] Y. Wang, H. Yan, Y. Zhang, P. Gao, F. Yu, X. Xiong, et al., “PSCaps: High-Performance Pose-Sensitive Layout Hotspot Detector based on CapsNet,” ACM Trans. Des. Autom. Electron. Syst., vol. 30, no. 4, pp. 1–21, 2025. 10.1145/3735132 DOI: https://doi.org/10.1145/3735132

[36] R. LaLonde, N. Khosravan, and U. Bagci, “Deformable capsules for object detection,” Adv. Intell. Syst., vol. 6, no. 9, p. 2400044, Sep. 2024. https://doi.org/10.1002/aisy.202400044 DOI: https://doi.org/10.1002/aisy.202400044

[37] M. K. Meesala, R. Vallabhaneni, M. Mathapati, P. K. Pareek, and J. Metan, 2024, September. Beyond the Horizon: Drone-Assisted HAR Through Cutting-Edge Caps Net and Optimization Techniques. In 2024 International Conference on Distributed Systems, Computer Networks and Cybersecurity (ICDSCNC) (pp. 1-6). IEEE. https://doi.org/10.1109/ICDSCNC62492.2024.10941663 DOI: https://doi.org/10.1109/ICDSCNC62492.2024.10941663

[38] L. De Laurentiis, A. Pomente, F. Del Frate, and G. Schiavon, “Capsule and convolutional neural network-based SAR ship classification in Sentinel-1 data,” in Active and Passive Microwave Remote Sensing for Environmental Monitoring III, vol. 11154. SPIE, 2019, October, p. 1115405. https://doi.org/10.1117/12.2532551 DOI: https://doi.org/10.1117/12.2532551

[39] A. Ergasheva, F. Akhmedov, A. Abdusalomov, and W. Kim, “Advancing Maritime Safety: Early Detection of Ship Fires through Computer Vision, Deep Learning Approaches, and Histogram Equalization Techniques,” Fire (Basel), vol. 7, no. 3, p. 84, 2024. DOI: https://doi.org/10.3390/fire7030084

[40] A. Abdusalomov, S. Umirzakova, A. Kutlimuratov, D. Mirzaev, A. Dauletov, T. Botirov, et al., “Lightweight UAV-Based System for Early Fire-Risk Identification in Wild Forests,” Fire (Basel), vol. 8, no. 8, p. 288, 2025. DOI: https://doi.org/10.3390/fire8080288

[41] R. Nasimov, A. Abdusalomov, C. Khidirova, K. Temirova, A. Kutlimuratov, S. Sadikova, et al., “PatientEase-Domain-Aware RAG for Rehabilitation Instruction Simplification,” Bioengineering (Basel), vol. 12, no. 11, p. 1204, Nov. 3 2025. DOI: https://doi.org/10.3390/bioengineering12111204

[42] F. Akhmedov, R. Nasimov, and A. Abdusalomov, “Developing a Comprehensive Oil Spill Detection Model for Marine Environments,” Remote Sens. (Basel), vol. 16, no. 16, p. 3080, 2024. DOI: https://doi.org/10.3390/rs16163080

[43] N. Yunusov, B. M. S. Islam, A. Abdusalomov, and W. Kim, “Robust Forest Fire Detection Method for Surveillance Systems Based on You Only Look Once Version 8 and Transfer Learning Approaches,” Processes (Basel), vol. 12, no. 5, p. 1039, 2024. DOI: https://doi.org/10.3390/pr12051039