|
|
|
| Small-scale sparse smoke detection in multiple fire scenarios based on semi-supervised learning |
Kaibo YANG1( ),Mingen ZHONG1,*(),Jiawei TAN2,Zhiying DENG1,Mengli ZHOU1,Ziji XIAO1 |
1. Fujian Key Laboratory of Bus Advanced Design and Manufacture, Xiamen University of Technology, Xiamen 361024, China
2. School of Aerospace Engineering, Xiamen University, Xiamen 361102, China |
|
|
|
Abstract A dataset named MSIFSD, which contained 9 types of fire scenes and 3 types of smoke, was constructed to explore a lightweight fire smoke detection algorithm with good performance. Additionally, a deep convolutional neural network named DeepSmoke was developed. To address the challenge of detecting small-scale and sparse smoke, an efficient feature aggregation module called PM-C2f was proposed and incorporated with a partial mixed-relevance region self-attention mechanism module called PM-TF. The PM-C2f module was utilized to integrate contextual information from various levels of image features, while the PM-TF module was used to strengthen the sparse features of small-scale and sparse smoke. A semi-supervised training strategy using a pseudo-label classifier was proposed to address the issue of DeepSmoke’s insufficient adaptability across different scenarios. A large amount of unlabeled data were leveraged to assist model training and improve the detection performance across multiple scene types. Experimental results demonstrated that the proposed algorithm achieved the detection accuracies of 88.2%, 90.0%, and 98.3% for small-scale, sparse, and general smoke, respectively. The average comprehensive detection accuracy was 94.2%, outperforming the existing mainstream algorithms. And the floating-point operation was 9.3×109, reflecting the friendliness to edge devices.
|
|
Received: 03 February 2024
Published: 10 March 2025
|
|
|
| Fund: 福建省自然科学基金资助项目(2023J011439, 2019J01859). |
|
Corresponding Authors:
Mingen ZHONG
E-mail: kaiboo_yang.ty@foxmail.com;zhongmingen@xmut.edu.cn
|
基于半监督学习的多场景火灾小规模稀薄烟雾检测
为了探索高性能的轻量级火灾烟雾检测算法，构建了包含9种火灾场景、3个烟雾类型的图像集MSIFSD，设计了深度卷积神经网络DeepSmoke. 针对小规模稀薄烟雾检测困难的问题，提出高效特征聚合模块PM-C2f和部分混合最相关区域自注意力机制模块PM-TF，PM-C2f模块用来整合各层次图像特征的上下文信息，PM-TF模块用于强化小规模稀薄烟雾的稀疏特征. 针对DeepSmoke在不同场景下适应性不足的问题，提出使用伪标签分类器的半监督训练策略，利用大量未标注数据辅助模型训练，提升多类场景下的检测性能. 实验结果表明，该算法针对小规模、稀薄烟雾和普通烟雾的检测准确率分别为88.2%、90.0%和98.3%，综合平均检测准确率为94.2%，均优于现有主流算法，且浮点运算量仅为9.3×109，体现了对边缘设备的友好性.
关键词：
火灾,
烟雾检测,
深度学习,
半监督学习,
伪标签,
注意力
|
|
| [1] |
FRIZZI S, KAABI R, BOUCHOUICHA M, et al. Convolutional neural network for video fire and smoke detection [C]// 42nd Annual Conference of the IEEE Industrial Electronics Society . Florence: IEEE, 2016: 877–882.
|
|
|
| [2] |
王殿伟, 赵文博, 房杰, 等 多类场景下无人机航拍视频烟雾检测算法[J]. 哈尔滨工业大学学报, 2023, 55 (10): 122- 129
WANG Dianwei, ZHAO Wenbo, FANG Jie, et al Smoke detection algorithm for UAV aerial video in multiple scenarios[J]. Journal of Harbin Institute of Technology, 2023, 55 (10): 122- 129
doi: 10.11918/202205119
|
|
|
| [3] |
马庆禄, 鲁佳萍, 唐小垚, 等 改进YOLOv5s的公路隧道烟火检测方法[J]. 浙江大学学报: 工学版, 2023, 57 (4): 784- 794
MA Qinglu, LU Jiaping, TANG Xiaoyao, et al Improved YOLOv5s flame and smoke detection method in road tunnels[J]. Journal of Zhejiang University: Engineering Science, 2023, 57 (4): 784- 794
|
|
|
| [4] |
谢康康, 朱文忠, 谢林森, 等 基于改进YOLOv7的火焰烟雾检测算法[J]. 国外电子测量技术, 2023, 42 (7): 41- 49
XIE Kangkang, ZHU Wenzhong, XIE Linsen, et al Improved YOLOv7-based flame smoke detection algorithm[J]. Foreign Electronic Measurement Technology, 2023, 42 (7): 41- 49
|
|
|
| [5] |
王晨灿, 李明 基于YOLOv8的火灾烟雾检测算法研究[J]. 北京联合大学学报, 2023, 37 (5): 69- 77
WANG Chencan, LI Ming Research on fire smoke detection algorithms based on YOLOv8[J]. Journal of Beijing Union University, 2023, 37 (5): 69- 77
|
|
|
| [6] |
HUO Y, ZHANG Q, JIA Y, et al A deep separable convolutional neural network for multiscale image-based smoke detection[J]. Fire Technology, 2022, 58 (7): 1445- 1468
|
|
|
| [7] |
WANG Y, HUA C, DING W, et al Real-time detection of flame and smoke using an improved YOLOv4 network[J]. Signal, Image and Video Processing, 2022, 16 (4): 1109- 1116
doi: 10.1007/s11760-021-02060-8
|
|
|
| [8] |
金程拓, 赵永智, 王涛, 等 基于改进YOLOX轻量级的烟雾火焰目标检测方法[J]. 滁州学院学报, 2023, 25 (5): 40- 45
JIN Chengtuo, ZHAO Yongzhi, WANG Tao, et al Improved YOLOX-based lightweight fire object detectiion algorithm[J]. Journal of Chuzhou University, 2023, 25 (5): 40- 45
doi: 10.3969/j.issn.1673-1794.2023.05.008
|
|
|
| [9] |
屠恩美, 杨杰 半监督学习理论及其研究进展概述[J]. 上海交通大学学报, 2018, 52 (10): 1280- 1291
TU Enmei, YANG Jie A review of semi-supervised learning theories and recent advances[J]. Journal of Shanghai Jiaotong University, 2018, 52 (10): 1280- 1291
|
|
|
| [10] |
XU B, CHEN M, GUAN W, et al. Efficient teacher: semi-supervised object detection for YOLOv5 [EB/OL]. (2023-02-15). https://arxiv.org/pdf/2302.07577.pdf.
|
|
|
| [11] |
周明非, 汪西莉 弱监督深层神经网络遥感图像目标检测模型[J]. 中国科学: 信息科学, 2018, 48 (8): 1022- 1034
ZHOU Mingfei, WANG Xili Object detection models of remote sensing images using deep neural networks with weakly supervised training method[J]. Scientia Sinica: Informationis, 2018, 48 (8): 1022- 1034
|
|
|
| [12] |
LIU Y C, MA C Y, HE Z, et al. Unbiased teacher for semi-supervised object detection [EB/OL]. (2021-02-18). https: //arxiv. org/abs/2102.09480. pdf.
|
|
|
| [13] |
ZHANG Y, YAO X, LIU C, et al. S4OD: semi-supervised learning for single-stage object detection [EB/OL]. [2023-12-16]. https://arxiv.org/abs/2204.04492.pdf.
|
|
|
| [14] |
XU M, ZHANG Z, HU H, et al. End-to-end semi-supervised object detection with soft teacher [C]// Proceedings of the IEEE/CVF International Conference on Computer Vision . Montreal: IEEE, 2021: 3060–3069.
|
|
|
| [15] |
CHEN B, LI P, CHEN X, et al. Dense learning based semi-supervised object detection [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . New Orleans: IEEE, 2022: 4815–4824.
|
|
|
| [16] |
MEETHAL A, PEDERSOLI M, ZHU Z, et al. Semi-weakly supervised object detection by sampling pseudo ground-truth boxes [EB/OL]. (2022-07-26). https://arxiv.org/abs/2204.00147.pdf.
|
|
|
| [17] |
YANG Q, WEI X, WANG B, et al. Interactive self-training with mean teachers for semi-supervised object detection [C]// Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . Nashville: IEEE, 2021: 5941–5950.
|
|
|
| [18] |
LIN T Y, DOLLAR P, GRISHICK R, et al. Feature pyramid networks for object detection [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition . Xiamen: IEEE, 2017: 2117–2125.
|
|
|
| [19] |
LIU S, QI L, QIN H, et al. Path aggregation network for instance segmentation [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition . Salt Lake City: IEEE, 2018: 8759–8768.
|
|
|
| [20] |
HE K, ZHANG X, REN S, et al Spatial pyramid pooling in deep convolutional networks for visual recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37 (9): 1904- 1916
doi: 10.1109/TPAMI.2015.2389824
|
|
|
| [21] |
WANG C Y, BOCHKOVSKIY A, LIAO H Y. YOLOv8: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors [C]// IEEE/CVF Conference on Computer Vision and Pattern Recognition . Vancouver: IEEE, 2023: 7464–7475.
|
|
|
| [22] |
CHEN J, KAO S, HE H, et al. Run, don’t walk: chasing higher flops for faster neural networks [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Vancouver: IEEE, 2023: 12021–12031.
|
|
|
| [23] |
LIU Y, SHAO Z, HOFFMANN N. Global attention mechanism: retain information to enhance channel-spatial interactions [EB/OL]. (2021-12-10). https://arxiv.org/pdf/2112.05561v1.pdf.
|
|
|
| [24] |
KLINKER F Exponential moving average versus moving exponential average[J]. Mathematische Semesterberichte, 2011, 58: 97- 107
doi: 10.1007/s00591-010-0080-8
|
|
|
| [25] |
ZHAN J, HU Y, ZHOU G, et al A high-precision forest fire smoke detection approach based on ARGNet[J]. Computers and Electronics in Agriculture, 2022, 196: 106874
doi: 10.1016/j.compag.2022.106874
|
|
|
| [26] |
YUAN F, ZHANG L, XIA X, et al A gated recurrent network with dual classification assistance for smoke semantic segmentation[J]. IEEE Transactions on Image Processing, 2021, 30: 4409- 442
doi: 10.1109/TIP.2021.3069318
|
|
|
| [27] |
PEDRO V, ADRIANO C, ADRIANO V An automatic fire detection system based on deep convolutional neural networks for low-power, resource-constrained devices[J]. Neural Computing and Applications, 2022, 30: 15349- 15368
|
|
|
| [28] |
JIE H, LI S, ALBANIE S, et al. Squeeze-and-excitation networks. [EB/OL]. (2019-05-16). https://arxiv.org/pdf/1709.01507.pdf.
|
|
|
| [29] |
WOO S, PARK J, LEE J, et al. CBAM: convolutional block attention module. [EB/OL]. (2018-07-18). https://arxiv.org/abs/1807.06521.
|
|
|
| [30] |
REN S, HE K, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks. [EB/OL]. (2016-01-06). https://arxiv.org/pdf/1506.01497.pdf.
|
|
|
| [31] |
LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector. [EB/OL]. (2015-12-08). https://arxiv.org/abs/1512.02325.
|
|
|
| [32] |
QIAO S, CHEN L, ALAN Y. DetectoRS: detecting objects with recursive feature pyramid and switchable atrous convolution. [EB/OL]. (2020-06-03). https://arxiv.org/abs/2006.02334.pdf.
|
|
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
| |
Shared |
|
|
|
|
| |
Discussed |
|
|
|
|
