《计算机应用》唯一官方网站 ›› 2025, Vol. 45 ›› Issue (4): 1317-1324.DOI: 10.11772/j.issn.1001-9081.2024040527
收稿日期:
2024-05-06
修回日期:
2024-10-26
接受日期:
2024-10-30
发布日期:
2025-01-03
出版日期:
2025-04-10
通讯作者:
张晓博
作者简介:
侯阳(1987—),女,黑龙江加格达奇人,工程师,硕士,主要研究方向:物联网(工业自动化方向)、大数据基金资助:
Yang HOU1, Qiong ZHANG2, Zixuan ZHAO2, Zhengyu ZHU2, Xiaobo ZHANG2()
Received:
2024-05-06
Revised:
2024-10-26
Accepted:
2024-10-30
Online:
2025-01-03
Published:
2025-04-10
Contact:
Xiaobo ZHANG
About author:
HOU Yang, born in 1987, M. S., engineer. Her research interests include internet of things (industrial automation direction), big data.Supported by:
摘要:
现有的烟火检测方法主要依赖员工现场巡视,效率低且实时性差,因此,提出一种基于YOLOv5s的复杂场景下的高效烟火检测算法YOLOv5s-MRD (YOLOv5s-MPDIoU-RevCol-Dyhead)。首先,采用MPDIoU (Maximized Position-Dependent Intersection over Union)方法改进边框损失函数,以适应重叠或非重叠的边界框回归(BBR),从而提高BBR的准确性和效率;其次,利用可逆柱状结构RevCol(Reversible Column)网络模型思想重构YOLOv5s模型的主干网络,使它具有多柱状网络架构,并在模型的不同层之间加入可逆链接,从而最大限度地保持特征信息以提高网络的特征提取能力;最后,引入Dynamic head检测头,以统一尺度感知、空间感知和任务感知,从而在不额外增加计算开销的条件下显著提高目标检测头的准确性和有效性。实验结果表明:在DFS(Data of Fire and Smoke)数据集上,与原始YOLOv5s算法相比,所提算法的平均精度均值(mAP@0.5)提升了9.3%,预测准确率提升了6.6%,召回率提升了13.8%。可见,所提算法能满足当前烟火检测应用场景的要求。
中图分类号:
侯阳, 张琼, 赵紫煊, 朱正宇, 张晓博. 基于YOLOv5s的复杂场景下高效烟火检测算法YOLOv5s-MRD[J]. 计算机应用, 2025, 45(4): 1317-1324.
Yang HOU, Qiong ZHANG, Zixuan ZHAO, Zhengyu ZHU, Xiaobo ZHANG. YOLOv5s-MRD: efficient fire and smoke detection algorithm for complex scenarios based on YOLOv5s[J]. Journal of Computer Applications, 2025, 45(4): 1317-1324.
实验 轮次 | MPDIoU | RevCol | Dyhead | 计算量/ GFLOPs | mAP@0.5 | 准确率 | 召回率 |
---|---|---|---|---|---|---|---|
1 | 24.00 | 0.43 | 0.91 | 0.80 | |||
2 | √ | 23.80 | 0.44 | 0.91 | 0.83 | ||
3 | √ | √ | 17.72 | 0.45 | 0.94 | 0.85 | |
4 | √ | √ | √ | 17.61 | 0.47 | 0.97 | 0.91 |
表1 不同改进方法对模型检测性能的影响
Tab. 1 Influence of different improved methods on model detection performance
实验 轮次 | MPDIoU | RevCol | Dyhead | 计算量/ GFLOPs | mAP@0.5 | 准确率 | 召回率 |
---|---|---|---|---|---|---|---|
1 | 24.00 | 0.43 | 0.91 | 0.80 | |||
2 | √ | 23.80 | 0.44 | 0.91 | 0.83 | ||
3 | √ | √ | 17.72 | 0.45 | 0.94 | 0.85 | |
4 | √ | √ | √ | 17.61 | 0.47 | 0.97 | 0.91 |
算法 | 准确率 | 召回率 |
---|---|---|
BoWFire | 0.90 | 0.66 |
Xception | 0.90 | 0.86 |
KutralNet | 0.92 | 0.67 |
FireNet | 0.93 | 0.90 |
YOLOv6s | 0.92 | 0.84 |
YOLOv8s | 0.93 | 0.85 |
YOLOv9s | 0.96 | 0.90 |
YOLOv5s-MRD | 0.98 | 0.91 |
表2 DFS数据集上的各算法的对比结果
Tab. 2 Comparison results of different algorithms on DFS dataset
算法 | 准确率 | 召回率 |
---|---|---|
BoWFire | 0.90 | 0.66 |
Xception | 0.90 | 0.86 |
KutralNet | 0.92 | 0.67 |
FireNet | 0.93 | 0.90 |
YOLOv6s | 0.92 | 0.84 |
YOLOv8s | 0.93 | 0.85 |
YOLOv9s | 0.96 | 0.90 |
YOLOv5s-MRD | 0.98 | 0.91 |
算法 | Fire Detection | D-Fire | ||
---|---|---|---|---|
准确率 | 召回率 | 准确率 | 召回率 | |
BoWFire | 0.90 | 0.57 | 0.79 | 0.61 |
Xception | 0.85 | 0.80 | 0.89 | 0.95 |
KutralNet | 0.93 | 0.92 | 0.85 | 0.85 |
FireNet | 0.92 | 0.86 | 0.87 | 0.89 |
YOLOv6s | 0.92 | 0.90 | 0.89 | 0.85 |
YOLOv8s | 0.93 | 0.92 | 0.88 | 0.88 |
YOLOv9s | 0.91 | 0.87 | 0.84 | 0.70 |
YOLOv5s-MRD | 0.94 | 0.93 | 0.99 | 0.83 |
表3 Fire Detection和D-Fire数据集上各算法的对比结果
Tab. 3 Comparison results of different algorithms on Fire Detection and D-Fire datasets
算法 | Fire Detection | D-Fire | ||
---|---|---|---|---|
准确率 | 召回率 | 准确率 | 召回率 | |
BoWFire | 0.90 | 0.57 | 0.79 | 0.61 |
Xception | 0.85 | 0.80 | 0.89 | 0.95 |
KutralNet | 0.93 | 0.92 | 0.85 | 0.85 |
FireNet | 0.92 | 0.86 | 0.87 | 0.89 |
YOLOv6s | 0.92 | 0.90 | 0.89 | 0.85 |
YOLOv8s | 0.93 | 0.92 | 0.88 | 0.88 |
YOLOv9s | 0.91 | 0.87 | 0.84 | 0.70 |
YOLOv5s-MRD | 0.94 | 0.93 | 0.99 | 0.83 |
模型 | 计算量/GFLOPs |
---|---|
YOLOv6s | 44.90 |
YOLOv8s | 28.82 |
YOLOv9s | 39.64 |
YOLOv5s-MRD | 17.61 |
表4 时间复杂度对比
Tab. 4 Comparison of time complexity
模型 | 计算量/GFLOPs |
---|---|
YOLOv6s | 44.90 |
YOLOv8s | 28.82 |
YOLOv9s | 39.64 |
YOLOv5s-MRD | 17.61 |
1 | DE LEÓN-RUIZ J E, CARVAJAL-MARISCAL I, DE LA CRUZ-ÁVILA M, et al. Image convolution-based experimental technique for flame front detection and dimension estimation: a case study on laminar-to-transition jet diffusion flame height measurement[J]. Measurement Science and Technology, 2022, 33(7): No.075406. |
2 | JEON M, CHOI H S, LEE J, et al. Multi-scale prediction for fire detection using convolutional neural network[J]. Fire Technology, 2021, 57(5):2533-2551. |
3 | BAI X, WANG Z. Research on forest fire detection technology based on deep learning[C]// Proceedings of the 2021 International Conference on Computer Network, Electronic and Automation. Piscataway: IEEE, 2021: 85-90. |
4 | WANG Z, WU L, LI T, et al. A smoke detection model based on improved YOLOv5[J]. Mathematics, 2022, 10(7): No.1190. |
5 | 吴凡. 基于深度学习的火灾检测算法研究与实现[D]. 杭州:杭州电子科技大学, 2020:1-67. |
WU F. Research and implementation of fire detection algorithm based on deep learning[D]. Hangzhou: Hangzhou Dianzi University, 2020:1-67. | |
6 | ZHOU X, WANG D, KRÄHENBÜHL P. Objects as points[EB/OL]. [2024-04-25].. |
7 | 谢书翰,张文柱,程鹏,等. 嵌入通道注意力的YOLOv4火灾烟雾检测模型[J]. 液晶与显示, 2021, 36(10): 1445-1453. |
XIE S H, ZHANG W Z, CHENG P, et al. Firesmoke detection model based on YOLOv4 with channel attention[J]. Chinese Journal of Liquid Crystals and Displays, 2021, 36(10): 1445-1453. | |
8 | 邹辉军,焦良葆,孟琳,等. 基于CG-yolo的烟火检测[J]. 计算机与数字工程, 2022, 50(1): 206-212. |
ZOU H J, JIAO L B, MENG L, et al. Detection of fireworks based on CG-yolo[J]. Computer and Digital Engineering, 2022, 50(1): 206-212. | |
9 | HE Y, HU J, ZENG M, et al. DCGC-YOLO: the efficient dual-channel bottleneck structure YOLO detection algorithm for fire detection[J]. IEEE Access, 2024, 12: 65254-65265. |
10 | WANG Y, HU Y. Smoke and fire detection algorithm based on improved YOLOv5[C]// Proceedings of the 6th International Conference on Intelligent Autonomous Systems. Piscataway: IEEE, 2023: 82-87. |
11 | XU S, JI Y, WANG G, et al. GFSPP-YOLO: a light YOLO model based on group fast spatial pyramid pooling[C]// Proceedings of the IEEE 11th International Conference on Information, Communication and Networks. Piscataway: IEEE, 2023: 733-738. |
12 | PHAN D T, YAP K H, GARG K, et al. Vision-based early fire and smoke detection for smart factory applications using FFS-YOLO[C]// Proceedings of the IEEE 25th International Workshop on Multimedia Signal Processing. Piscataway: IEEE, 2023: 1-6. |
13 | WANG X, JING Z, SHI L, et al. MA-YOLO: a lightweight vehicle detection framework based on YOLO[C]// Proceedings of the 2023 International Conference on Artificial Intelligence and Automation Control. Piscataway: IEEE, 2023: 277-281. |
14 | CAI Y, ZHOU Y, HAN Q, et al. Reversible column networks[EB/OL]. [2023-02-01].. |
15 | MA S, XU Y. MPDIoU: a loss for efficient and accurate bounding box regression[EB/OL]. [2023-07-14].. |
16 | DAI X, CHEN Y, XIAO B, et al. Dynamic head: unifying object detection heads with attentions[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021: 7369-7378. |
17 | DAI J, QI H, XIONG Y, et al. Deformable convolutional networks[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2017: 764-773. |
18 | GIRSHICK R. Fast R-CNN[C]// Proceedings of the 2015 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2015: 1440-1448. |
19 | HE K, GKIOXARI G, DOLLÁR P, et al. Mask R-CNN[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2017: 2980-2988. |
20 | REN S, HE K, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence. Piscataway: IEEE, 2017, 39(6): 1137-1149. |
21 | TIAN Z, SHEN C, CHEN H, et al. FCOS: fully convolutional one-stage object detection[C]// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2019: 9626-9635. |
22 | WANG X, ZHANG S, YU Z, et al. Scale-equalizing pyramid convolution for object detection[C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 13356-13365. |
23 | YANG Z, LIU S, HU H, et al. RepPoints: point set representation for object detection[C]// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2019: 9656-9665. |
24 | ZHANG S, CHI C, YAO Y, et al. Bridging the gap between anchor-based and anchor-free detection via adaptive training sample selection[C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 9756-9765. |
25 | CHEN Y, DAI X, LIU M, et al. Dynamic ReLU[C]// Proceedings of the 2020 European Conference on Computer Vision, LNCS 12364. Cham: Springer, 2020: 351-367. |
26 | DUNNINGS A J, BRECKON T P. Experimentally defined convolutional neural network architecture variants for non-temporal real-time fire detection[C]// Proceedings of the 25th IEEE International Conference on Image Processing. Piscataway: IEEE, 2018: 1558-1562. |
27 | SHAMSOSHOARA A, AFGHAH F, RAZI A, et al. Aerial imagery pile burn detection using deep learning: the FLAME dataset[J]. Computer Networks, 2021, 193: No.108001. |
28 | AYALA A, FERNANDES B, CRUZ F, et al. KutralNet: a portable deep learning model for fire recognition[C]// Proceedings of the 2020 International Joint Conference on Neural Networks. Piscataway: IEEE, 2020: 1-8. |
29 | LI C, LI L, JIANG H, et al. YOLOv6: a single-stage object detection framework for industrial applications[EB/OL]. [2024-09-07].. |
30 | Ultralytics. YOLOv8[EB/OL]. [2024-04-26].. |
31 | WANG C Y, YEH I H, LIAO H Y M. YOLOv9: learning what you want to learn using programmable gradient information[C]// Proceedings of the 2024 European Conference on Computer Vision, LNCS 15089. Cham: Springer, 2025: 1-21. |
32 | CHINO D Y T, AVALHAIS L P S, RODRIGUES J F, et al. BoWFire: detection of fire in still images by integrating pixel color and texture analysis[C]// Proceedings of the 28th SIBGRAPI Conference on Graphics, Patterns and Images. Piscataway: IEEE, 2015: 95-102. |
[1] | 张李伟, 梁泉, 胡禹涛, 朱乔乐. 基于分组卷积的通道重洗注意力机制[J]. 《计算机应用》唯一官方网站, 2025, 45(4): 1069-1076. |
[2] | 赵轻轻, 胡滨. 不变性全局稀疏轮廓点表征的运动行人检测神经网络[J]. 《计算机应用》唯一官方网站, 2025, 45(4): 1271-1284. |
[3] | 张传浩, 屠晓涵, 谷学汇, 轩波. 基于多模态信息相互引导补充的雷达-相机三维目标检测[J]. 《计算机应用》唯一官方网站, 2025, 45(3): 946-952. |
[4] | 余松森, 林智凡, 薛国鹏, 徐建宇. 基于改进YOLOv8的轻量级大幅面瓷砖缺陷检测算法[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 647-654. |
[5] | 洪梓榕, 包广清. 基于集成学习的雷达自动目标识别综述[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 371-382. |
[6] | 杨晟, 李岩. 面向目标检测的对比知识蒸馏方法[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 354-361. |
[7] | 桂佳扬, 王顺吉, 周正康, 唐加山. 基于改进YOLOv8n的隧道内异物检测算法[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 655-661. |
[8] | 文诗佳, 金世俊. 结合目标检测和特征点关联的动态视觉SLAM算法[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 610-615. |
[9] | 张众维, 王俊, 刘树东, 王志恒. 多尺度特征融合与加权框融合的遥感图像目标检测[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 633-639. |
[10] | 何秋润, 胡节, 彭博, 李天源. 基于上下文信息的多尺度特征融合织物疵点检测算法[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 640-646. |
[11] | 杨博然, 蔺素珍, 李大威, 禄晓飞, 崔晨辉. 基于信息补偿的红外弱小目标检测方法[J]. 《计算机应用》唯一官方网站, 2025, 45(1): 284-291. |
[12] | 刘赏, 周煜炜, 代娆, 董林芳, 刘猛. 融合注意力和上下文信息的遥感图像小目标检测算法[J]. 《计算机应用》唯一官方网站, 2025, 45(1): 292-300. |
[13] | 潘烨新, 杨哲. 基于多级特征双向融合的小目标检测优化模型[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2871-2877. |
[14] | 李烨恒, 罗光圣, 苏前敏. 基于改进YOLOv5的Logo检测算法[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2580-2587. |
[15] | 张英俊, 李牛牛, 谢斌红, 张睿, 陆望东. 课程学习指导下的半监督目标检测框架[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2326-2333. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||