An Improvement of the Fire Detection and Classification Method Using YOLOv3 for Surveillance Systems
Open accessAbstract
Currently, sensor-based systems for fire detection are widely used worldwide. Further research has shown that camera-based fire detection systems achieve much better results than sensor-based methods. In this study, we present a method for real-time high-speed fire detection using deep learning. A new special convolutional neural network was developed to detect fire regions using the existing YOLOv3 algorithm. Due to the fact that our real-time fire detector cameras were built on a Banana Pi M3 board, we adapted the YOLOv3 network to the board level. Firstly, we tested the latest versions of YOLO algorithms to select the appropriate algorithm and used it in our study for fire detection. The default versions of the YOLO approach have very low accuracy after training and testing in fire detection cases. We selected the YOLOv3 network to improve and use it for the successful detection and warning of fire disasters. By modifying the algorithm, we recorded the results of a rapid and high-precision detection of fire, during both day and night, irrespective of the shape and size. Another advantage is that the algorithm is capable of detecting fires that are 1 m long and 0.3 m wide at a distance of 50 m. Experimental results showed that the proposed method successfully detected fire candidate areas and achieved a seamless classification performance compared to other conventional fire detection frameworks.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The modified YOLOv3 training approach allows algorithm to utilize dataset containing images of both flames and smoke, enabling it to recognize and differentiate between these two features.
“This training approach allows the algorithm to utilize a dataset that contains images of both flames and smoke, which enables it to effectively recognize and differentiate between these two features.”
From A Survey of Video Surveillance Systems in Smart City (2023) · cites this paper as [133]
Section 3.1 (Dataset; page unavailable) says the YOLOv3 training images were labeled with flame coordinates and included fire-like negative images, supporting the flame-detection part. It does not say the modified YOLOv3 model was trained to recognize or distinguish smoke. Section 2.2 discusses a fire-and-smoke dataset, but explicitly identifies it as part of the authors’ earlier CNN work, not this YOLOv3 approach. (LLM, conf=0.87)
Experimental — accuracy not yet measuredTraining modified YOLOv3 on images containing red color associated with fire improved accuracy of recognition method.
“The authors improved the accuracy of the recognition method by training the algorithm on images that specifically contained the red color that is typically associated with fire.”
From A Survey of Video Surveillance Systems in Smart City (2023) · cites this paper as [133]
Section 3.3 (page unavailable) reports that the improved YOLOv3 was trained on an expanded fire-image dataset and achieved higher accuracy. It also says retaining flame color mattered because color augmentation reduced accuracy, but it does not say the training images were selected specifically for their red color. (LLM, conf=0.86)
Experimental — accuracy not yet measuredTraining modified YOLOv3 on fire-related images makes model more adept at calculating probability of smoke appearing before flames visible.
“By training the modified YOLOv3 algorithm on fire-related images, the model becomes more adept at accurately calculating the probability of smoke appearing before any flames become visible.”
From A Survey of Video Surveillance Systems in Smart City (2023) · cites this paper as [133]
Section 3.2 (page not available) says the authors augmented fire images to train YOLOv3 and used the model to predict fire scores; it also reports early-stage detection of small fire regions. It does not say the model calculates the probability of smoke appearing before flames become visible. (LLM, conf=0.88)
Experimental — accuracy not yet measuredA modified YOLOv3 algorithm developed in [133] detects and classifies regions affected by fire with accuracy 98.9% in 2021.
“In [133], a modified version of the YOLOv3 algorithm was developed to detect and classify regions affected by fire with an accuracy of 98.9% in 2021.”
From A Survey of Video Surveillance Systems in Smart City (2023) · cites this paper as [133]
The Introduction and Conclusions describe an improved YOLOv3 method for detecting and classifying fire regions. However, Section 4 (Experimental Results and Discussion) identifies 98.9% as the average of F-measure, recall, and precision—not accuracy. Section 3.3 reports a final fire-detection accuracy of 99.7%. (LLM, conf=0.90)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · see citing works
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 74,248 · see citing works
Szeliski, R. (2010) · Texts in computer science · doi:10.1007/978-1-84882-935-0 · Cited by 1,409 · see citing works
A Forest Fire Detection System Based on Ensemble Learning
Xu, R., Lin, H., Lu, K., Cao, L., & Liu, Y. (2021) · Forests · doi:10.3390/f12020217 · Cited by 575 · see citing works
Convolutional Neural Networks Based Fire Detection in Surveillance Videos
Muhammad, K., Ahmad, J., Mehmood, I., Rho, S., & Baik, S. W. (2018) · IEEE Access · doi:10.1109/access.2018.2812835 · Cited by 556 · see citing works
A Review on Early Forest Fire Detection Systems Using Optical Remote Sensing
Barmpoutis, P., Papaioannou, P., Dimitropoulos, K., & Grammalidis, N. (2020) · Sensors · doi:10.3390/s20226442 · Cited by 547 · see citing works
Image fire detection algorithms based on convolutional neural networks
Li, P., & Zhao, W. (2020) · Case Studies in Thermal Engineering · doi:10.1016/j.csite.2020.100625 · Cited by 377 · see citing works
Fire detection using statistical color model in video sequences
Çelik, T., Demirel, H., Özkaramanlı, H., & Uyguroğlu, M. K. (2007) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2006.12.003 · Cited by 342 · see citing works
Spatio-Temporal Flame Modeling and Dynamic Texture Analysis for Automatic Video-Based Fire Detection
Dimitropoulos, K., Barmpoutis, P., & Grammalidis, N. (2014) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2014.2339592 · Cited by 264 · see citing works
A Video-Based Fire Detection Using Deep Learning Models
Kim, B., & Lee, J. (2019) · Applied Sciences · doi:10.3390/app9142862 · Cited by 206 · see citing works
Ba, R., Chen, C., Yuan, J. Y., Song, W., & Lo, S. (2019) · Remote Sensing · doi:10.3390/rs11141702 · Cited by 162 · see citing works
Fire smoke detection algorithm based on motion characteristic and convolutional neural networks
Luo, Y., Zhao, L., Liu, P., & Huang, D. (2017) · Multimedia Tools and Applications · doi:10.1007/s11042-017-5090-2 · Cited by 152 · see citing works
Using Popular Object Detection Methods for Real Time Forest Fire Detection
Wu, S., & Zhang, L. (2018) · International Symposium on Computational Intelligence and Design · doi:10.1109/iscid.2018.00070 · Cited by 152 · see citing works
Barmpoutis, P., Stathaki, T., Dimitropoulos, K., et al. (2020) · Remote Sensing · doi:10.3390/rs12193177 · Cited by 114 · see citing works
Efficient Flame Detection Based on Static and Dynamic Texture Analysis in Forest Fire Detection
Prema, C. E., Vinsley, S. S., & Suresh, S. (2017) · Fire Technology · doi:10.1007/s10694-017-0683-x · Cited by 111 · see citing works
Toulouse, T., Rossi, L., Çelik, T., et al. (2015) · Signal Image and Video Processing · doi:10.1007/s11760-015-0789-x · Cited by 93 · see citing works
Automatic Fire and Smoke Detection Method for Surveillance Systems Based on Dilated CNNs
Valikhujaev, Y., Abdusalomov, A., & Cho, Y. (2020) · Atmosphere · doi:10.3390/atmos11111241 · Cited by 92 · see citing works
Self-citationPark, M., & Ko, B. C. (2020) · Sensors · doi:10.3390/s20082202 · Cited by 80 · see citing works
Pan, H., Badawi, D., & Çetin, A. E. (2020) · Sensors · doi:10.3390/s20102891 · Cited by 78 · see citing works
Contour Based Forest Fire Detection Using FFT and Wavelet
Zhang, Z., Zhao, J., Zhang, D., Qu, C., Ke, Y., & Cai, B. (2008) · doi:10.1109/csse.2008.837 · Cited by 65 · see citing works
Detection of Wildfire Smoke Images Based on a Densely Dilated Convolutional Network
Li, T., Zhao, E., Zhang, J., & Hu, C. (2019) · Electronics · doi:10.3390/electronics8101131 · Cited by 59 · see citing works
Vision based monitoring and characterisation of combustion flames
Lu, G., Gilabert, G., & Yan, Y. G. (2005) · Journal of Physics Conference Series · doi:10.1088/1742-6596/15/1/033 · Cited by 58 · see citing works
Gagliardi, A., & Saponara, S. (2020) · Energies · doi:10.3390/en13082098 · Cited by 32 · see citing works
Large Space Fire Image Processing of Improving Canny Edge Detector Based on Adaptive Smoothing
Jiang, Q., & Wang, Q. (2010) · doi:10.1109/cicc-itoe.2010.73 · Cited by 31 · see citing works
A Fire Monitoring and Alarm System Based on YOLOv3 with OHEM
Shi, F., Qian, H., Chen, W., et al. (2020) · Chinese Control Conference (CCC) · doi:10.23919/ccc50068.2020.9189667 · Cited by 22 · see citing works
Study of Flame Detection based on Improved YOLOv4
Cao, C., Tan, X., Huang, X. L. D., et al. (2021) · Journal of Physics Conference Series · doi:10.1088/1742-6596/1952/2/022016 · Cited by 10 · see citing works
Lee, S. J., Kim, B. H., & Kim, M. Y. (2021) · IEEE Access · doi:10.1109/access.2021.3078623 · Cited by 7 · see citing works
Sharma, J., Granmo, O., & Goodwin, M. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-79457-6_9 · Cited by 4 · see citing works
Contiguous Loss for Motion-Based, Non-Aligned Image Deblurring
Niu, W., Xia, K., & Pan, Y. (2021) · Symmetry · doi:10.3390/sym13040630 · Cited by 3 · see citing works
Flagged first, then by citations
Research works printed but not linked: 2
- [25]Redmon J. Farhadi A. YOLOv3: An incremental improvement arXiv 2018 1804.02767
- [28]He K. Zhang X. Ren S. Sun J. Deep Residual Learning for Image Recognition Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Las Vegas, NV, USA 26 June–1 July 2016 770 778
Cited by
Sorted by citations
An improved fire detection approach based on YOLO-v8 for smart cities
Talaat, F. M., & ZainEldin, H. (2023) · Neural Computing and Applications · doi:10.1007/s00521-023-08809-1 · Cited by 642 · see citing works
A Survey of Video Surveillance Systems in Smart City
Myagmar-Ochir, Y., & Kim, W. (2023) · Electronics · doi:10.3390/electronics12173567 · Cited by 134 · see citing works
Wu, H., Hu, Y., Wang, W., Mei, X., & Xian, J. (2022) · Sensors · doi:10.3390/s22197420 · Cited by 63 · see citing works
Namana, M. S. K., & Kumar, B. U. (2024) · International Journal of Safety and Security Engineering · doi:10.18280/ijsse.140611 · Cited by 6
- Authors
- Abdusalomov, A., Baratov, N., Kutlimuratov, A., & Whangbo, T. K.
- Affiliations
- Gachon University (KR)
- Tashkent University of Information Technology (UZ)
- Corresponding author
- Whangbo, Taeg Keun
- Venue
- Sensors
- Publisher
- MDPI AG
- Published
- 2021-09-29
- Volume
- 21
- Issue
- 19
- Pages
- 6519
- ISSN
- 1424-8220 (ISSN-L)
- PMC
- PMC8511986
- OpenAlex
- W3202825385
- Cited by
- 131
- References
- 29
- Citations by year
- 2026: 10 · 2025: 30 · 2024: 34 · 2023: 33 · 2022: 22 · 2021: 2
- FWCI
- 17.02
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Sensors (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- National Research Foundation
- Topics
- Physical Sciences › Engineering › Safety, Risk, Reliability and Quality › Fire Detection and Safety Systems
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Engineering › Safety, Risk, Reliability and Quality › Fire dynamics and safety research
- Keywords
- fire detection, fire classification, YOLOv3, real-time object detection, video surveillance, edge computing, nighttime detection, fire alarm systems
- MeSH terms
- Algorithms, Fires, Neural Networks, Computer
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
