Firesmoke detection model based on YOLOv4 with channel attention

Shu-han XIE; Wen-zhu ZHANG; Peng CHEN; Zi-xuan YANG

doi:10.37188/CJLCD.2020-0312

您当前的位置：

首页 >

文章列表页 >

Firesmoke detection model based on YOLOv4 with channel attention

Image Processing | 更新时间：2021-09-30

- Firesmoke detection model based on YOLOv4 with channel attention
- Chinese Journal of Liquid Crystals and Displays Vol. 36, Issue 10, Pages: 1445-1453(2021)
- 作者机构：
  
  西安建筑科技大学信息与控制工程学院, 陕西西安 710048
- 作者简介：
- 基金信息：
  
  Natural Science Basic Research Program of Shanxi Province(2020JM-489);Xi′an Science and Technology Plan Project(JZKD0010);Shaanxi Provincial Department of Education Key Project High-end Think Tank(18JT006)
- DOI：10.37188/CJLCD.2020-0312
  CLC： TP391.4
- Received：20 November 2020，
  
  Revised：02 March 2021，
  
  Published：2021-10
- 稿件说明：
移动端阅览
Shu-han XIE, Wen-zhu ZHANG, Peng CHEN, et al. Firesmoke detection model based on YOLOv4 with channel attention[J]. Chinese journal of liquid crystals and displays, 2021, 36(10): 1445-1453.
DOI：

Shu-han XIE, Wen-zhu ZHANG, Peng CHEN, et al. Firesmoke detection model based on YOLOv4 with channel attention[J]. Chinese journal of liquid crystals and displays, 2021, 36(10): 1445-1453. DOI： 10.37188/CJLCD.2020-0312.

摘要

为提高火灾烟雾检测模型的性能，以及避免繁琐的手工烟雾特征提取过程，本文提出一种基于卷积神经网络YOLOv4的火灾烟雾检测模型。该方法以CSPDarkNet53为主干网络，在主干网络的最后一层加入了13×13.9×9.5×5.1×1四个不同尺度的最大池化，多尺度特征融合中采用了PANet（Path Aggregation Network）以提高网络特征提取能力；为了增强网络预测头提取有效烟雾信息的能力，在网络预测头加入通道注意力网络。针对火灾烟雾数据集，候选框的尺寸使用K-means进行聚类以得到更加符合贴近火灾烟雾数据集的尺寸；由于本文仅识别火灾烟雾一种类别，所以精简损失函数，剔除分类误差，使算法收敛得更快。训练阶段使用了图像翻转、放缩和随机擦除等数据增强方法以降低过拟合的风险。实验结果表明，所提出的火灾烟雾检测模型精度高，其准确率达到92.5%，召回率达到87.7%，同时检测速度可达5.1帧/s，提高了火灾烟雾检测模型的性能。

Abstract

To improve the precision and recall rate of fire smoke detection model in multi-scene fire smoke detection applications

and avoid the tedious manual smoke feature extraction process

a fire smoke detection model is proposed which is based on convolutional neural network YOLOv4. In the last layer of the backbone network

four different scales of maximum pooling are added: 13×13

9×9

5×5

and 1×1. The multi-scale feature fusion uses PANet (Path Aggregation Network) to improve network feature extraction capabilities. In addition

a channel attention network is added to the network prediction head to enhance the ability of the YOLO Head to extract effective smoke information. For the fire smoke data set

the size of the candidate frame is clustered using the K-means algorithm to get a size closer to the fire smoke data set. Due to the identification of smoke

the loss function is simplified

the classification error is eliminated

and the algorithm converges faster. Data enhancement methods such as image flipping and random erasure are used in the training phase to reduce the risk of overfitting. Experimental results show that the fire smoke detection model has excellent performance. Its precision can reach 92.5%

Recall can reach 87.7%

and the detection speed can reach 5.1 frames/s

which improves the performance of fire smoke detection model in multi-scene fire smoke detection applications.

关键词

Keywords

references

史劲亭 , 袁非牛 , 夏雪 . 视频烟雾检测研究进展 . 中国图象图形学报 , 2018 . 23 ( 3 ): 303 - 322 . https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201803001.htm https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201803001.htm .

J T SHI , F N YUAN , X XIA . Video smoke detection: a literature survey . Journal of Image and Graphics , 2018 . 23 ( 3 ): 303 - 322 . https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201803001.htm https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201803001.htm .

GENOVESE A, LABATI R D, PIURI V, et al . Wildfire smoke detection using computational intelligence techniques[C]// Proceedings of 2011 IEEE International Conference on Computational Intelligence for Measurement Systems and Applications . Ottawa, ON, Canada: IEEE, 2011: 1-6.

SEDLÁK V, STOPJAKOVÁ V, BRENKUŠ J. A real-time method for smoke detection in monitored forest areas[C]// Proceedings of 2017 International Conference on Applied Electronics . Pilsen, Czech Republic: IEEE, 2017: 1-6.

J GUBBI , S MARUSIC , M PALANISWAMI . Smoke detection in video using wavelets and support vector machines . Fire Safety Journal , 2009 . 44 ( 8 ): 1110 - 1115 . DOI: 10.1016/j.firesaf.2009.08.003 http://doi.org/10.1016/j.firesaf.2009.08.003 .

李红娣 , 袁非牛 . 采用金字塔纹理和边缘特征的图像烟雾检测 . 中国图象图形学报 , 2015 . 20 ( 6 ): 772 - 780 . https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201506006.htm https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201506006.htm .

H D LI , F N YUAN . Image based smoke detection using pyramid texture and edge features . Journal of Image and Graphics , 2015 . 20 ( 6 ): 772 - 780 . https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201506006.htm https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201506006.htm .

Q X ZHANG , G H LIN , Y M ZHANG , 等 . Wildland forest fire smoke detection based on faster R-CNN using synthetic smoke images . Procedia Engineering , 2018 . 211 441 - 446 . DOI: 10.1016/j.proeng.2017.12.034 http://doi.org/10.1016/j.proeng.2017.12.034 .

张倩 , 周平平 , 王公堂 , 等 . 基于合成图像的Faster R-CNN森林火灾烟雾检测 . 山东师范大学学报(自然科学版) , 2019 . 34 ( 2 ): 180 - 185 . DOI: 10.3969/j.issn.1001-4748.2019.02.010 http://doi.org/10.3969/j.issn.1001-4748.2019.02.010 .

Q ZHANG , P P ZHOU , G T WANG , 等 . Faster R-CNN forest fire smoke detection based on synthetic images . Journal of Shandong Normal University (Natural Science) , 2019 . 34 ( 2 ): 180 - 185 . DOI: 10.3969/j.issn.1001-4748.2019.02.010 http://doi.org/10.3969/j.issn.1001-4748.2019.02.010 .

FRIZZI S, KAABI R, BOUCHOUICHA M, et al . Convolutional neural network for video fire and smoke detection[C]// Proceedings of the 42 nd Annual Conference of the IEEE Industrial Electronics Society . Florence, Italy: IEEE, 2016: 877-882.

李鹏 , 张炎 . 基于高斯混合模型和卷积神经网络的视频烟雾检测 . 激光与光电子学进展 , 2019 . 56 ( 21 ): 211502 https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ201921019.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ201921019.htm .

P LI , Y ZHANG . Video smoke detection based on Gaussian mixture model and convolutional neural network . Laser & Optoelectronics Progress , 2019 . 56 ( 21 ): 211502 https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ201921019.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ201921019.htm .

富雅捷 , 张宏立 . 基于迁移学习的卷积神经网络森林火灾检测方法 . 激光与光电子学进展 , 2020 . 57 ( 4 ): 041010 https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202004013.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202004013.htm .

Y J FU , H L ZHANG . Forest fire detection method based on transfer learning of convolutional neural network . Laser & Optoelectronics Progress , 2020 . 57 ( 4 ): 041010 https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202004013.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202004013.htm .

冯路佳 , 王慧琴 , 王可 , 等 . 基于目标区域的卷积神经网络火灾烟雾识别 . 激光与光电子学进展 , 2020 . 57 ( 16 ): 161004 https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202016008.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202016008.htm .

L J FENG , H Q WANG , K WANG , 等 . Convolutional neural network fire smoke recognition based on target region . Laser & Optoelectronics Progress , 2020 . 57 ( 16 ): 161004 https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202016008.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202016008.htm .

刘丽娟 , 陈松楠 . 一种基于改进SSD的烟雾实时检测模型 . 信阳师范学院学报(自然科学版) , 2020 . 33 ( 2 ): 305 - 311 . DOI: 10.3969/j.issn.1003-0972.2020.02.021 http://doi.org/10.3969/j.issn.1003-0972.2020.02.021 .

L J LIU , S N CHEN . Real-time smoke detection model based on improved SSD . Journal of Xinyang Normal University (Natural Science Edition) , 2020 . 33 ( 2 ): 305 - 311 . DOI: 10.3969/j.issn.1003-0972.2020.02.021 http://doi.org/10.3969/j.issn.1003-0972.2020.02.021 .

BOCHKOVSKIY A, WANG C Y, LIAO H Y M. YOLOv4: optimal speed and accuracy of object detection[EB/OL]. (2020-04-23)[2020-05-20] . https://arxiv.org/abs/2004.10934 https://arxiv.org/abs/2004.10934 .

REDMON J, FARHADI A. Yolov3: an incremental improvement[EB/OL]. (2018-04-08)[2020-05-22] . https://arxiv.org/abs/1804.02767 https://arxiv.org/abs/1804.02767 .

董潇潇 , 何小海 , 吴晓红 , 等 . 基于注意力掩模融合的目标检测算法 . 液晶与显示 , 2019 . 34 ( 8 ): 825 - 833 . DOI: 10.3788/YJYXS20193408.0825 http://doi.org/10.3788/YJYXS20193408.0825 .

X X DONG , X H HE , X H WU , 等 . Object detection algorithm based on attention mask fusion . Chinese Journal of Liquid Crystals and Displays , 2019 . 34 ( 8 ): 825 - 833 . http://cjlcd.lightpublishing.cn/thesisDetails#10.3788/YJYXS20193408.0825&graphicAbstract=0 http://cjlcd.lightpublishing.cn/thesisDetails#10.3788/YJYXS20193408.0825&graphicAbstract=0 .

周鹏 , 姚剑敏 , 林志贤 , 等 . 融合注意力机制的移动端人像分割网络 . 液晶与显示 , 2020 . 35 ( 6 ): 547 - 554 . DOI: 10.3788/YJYXS20203506.0547 http://doi.org/10.3788/YJYXS20203506.0547 .

P ZHOU , J M YAO , Z X LIN , 等 . Mobile-based portrait segmentation network with attention mechanism . Chinese Journal of Liquid Crystals and Displays , 2020 . 35 ( 6 ): 547 - 554 . http://cjlcd.lightpublishing.cn/thesisDetails#10.3788/YJYXS20203506.0547&graphicAbstract=0 http://cjlcd.lightpublishing.cn/thesisDetails#10.3788/YJYXS20203506.0547&graphicAbstract=0 .

HU J, SHEN L, ALBANIE S, et al . Squeeze-and-excitation networks[EB/OL]. (2017-09-05)[2020-05-30] . https://arxiv.org/abs/1709.01507 https://arxiv.org/abs/1709.01507 .

张海涛 , 张梦 . 引入通道注意力机制的SSD目标检测算法 . 计算机工程 , 2020 . 46 ( 8 ): 264 - 270 . https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC202008039.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC202008039.htm .

H T ZHANG , M ZHANG . SSD target detection algorithm with channel attention mechanism . Computer Engineering , 2020 . 46 ( 8 ): 264 - 270 . https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC202008039.htm https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC202008039.htm .

李慕锴 , 张涛 , 崔文楠 . 基于YOLOv3的红外行人小目标检测技术研究 . 红外技术 , 2020 . 42 ( 2 ): 176 - 181 . https://www.cnki.com.cn/Article/CJFDTOTAL-HWJS202002014.htm https://www.cnki.com.cn/Article/CJFDTOTAL-HWJS202002014.htm .

M K LI , T ZHANG , W N CUI . Research of infrared small pedestrian target detection based on YOLOv3 . Infrared Technology , 2020 . 42 ( 2 ): 176 - 181 . https://www.cnki.com.cn/Article/CJFDTOTAL-HWJS202002014.htm https://www.cnki.com.cn/Article/CJFDTOTAL-HWJS202002014.htm .

孔祥勇. 基于深度学习的黑色素瘤分类问题的研究[D]. 杭州: 浙江工业大学, 2019.

KONG X Y. Research on classification of melanoma based on deep learning[D]. Hangzhou: Zhejiang University of Technology, 2019. (in Chinese)

金翠 , 王洪元 , 陈首兵 . 基于随机擦除行人对齐网络的行人重识别方法 . 山东大学学报(工学版) , 2018 . 48 ( 6 ): 67 - 73 . https://www.cnki.com.cn/Article/CJFDTOTAL-SDGY201806008.htm https://www.cnki.com.cn/Article/CJFDTOTAL-SDGY201806008.htm .

C JIN , H Y WANG , S B CHEN . Person re-identification based on random erasing pedestrian alignment network method . Journal of Shandong University (Engineering Science) , 2018 . 48 ( 6 ): 67 - 73 . https://www.cnki.com.cn/Article/CJFDTOTAL-SDGY201806008.htm https://www.cnki.com.cn/Article/CJFDTOTAL-SDGY201806008.htm .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Improved algorithm of GDT-YOLOV 3 image target detection

Related Author

Shu-han XIE

Wen-zhu ZHANG

Peng CHEN

Zi-xuan YANG

Yue TANG

Ge WU

Yan PIAO

Related Institution

School of Information and Control Engineering， Xi'an University of Architecture and Technology

School of Electronics and Information Engineering, Changchun University of Science and Technology

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China 130033 Postal code：130033
Tel：0431-86176059 Email：yjxs@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰