基于改进双流卷积网络的火灾图像特征提取方法

doi:10.11896/jsjkx.180901640

计算机科学 ›› 2019, Vol. 46 ›› Issue (11): 291-296.doi: 10.11896/jsjkx.180901640

基于改进双流卷积网络的火灾图像特征提取方法

徐登^1,2, 黄晓东³

(江苏省物联网与制造业信息化工程技术研究开发中心江苏常州213164)¹
(常州机电职业技术学院信息工程学院江苏常州 213164)²
(东南大学MEMS教育部重点实验室南京210096)³

收稿日期:2018-09-27 出版日期:2019-11-15 发布日期:2019-11-14
通讯作者: 徐登(1982-),男,硕士,讲师,主要研究方向为嵌入式技术应用、图像处理等,E-mail:18720981239@163.com
作者简介:黄晓东(1982-),男,博士,副教授,主要研究方向为传感器及嵌入式技术应用。
基金资助:
本文受国家自然科学基金项目(61604040)资助。

Fire Images Features Extraction Based on Improved Two-stream Convolution Network

XU Deng^1,2, HUANG Xiao-dong³

(Jiangsu Internet of Things and Manufacturing Information Engineering Research Center,Changzhou,Jiangsu 213164,China)¹
(School of Information Engineering,Changzhou Vocational Institute of Mechatronic Technology,Changzhou,Jiangsu 213164,China)²
(Key Laboratory of MEMS of Ministry of Education,Southeast University,Nanjing 210096,China)³

Received:2018-09-27 Online:2019-11-15 Published:2019-11-14

摘要/Abstract

摘要： 基于图像处理技术的火灾监测,是近年来火灾监控领域的重要分支。对于开阔场景的火灾监测,利用火灾发生时产生的烟雾和火焰的动、静特性,以双流(Two-Stream)卷积神经网络作为理论基础对火灾进行检测识别。双流卷积神经网络采用空间流与时序流分别提取视频中的空间信息与时序信息,然而火灾初期的信息较为微弱,特征不够明显。为进一步提高初期的识别率,提出一种空间增强网络作为双流卷积神经网络的空间流来提取并增强视频的空间信息。空间增强网络同时对当前帧图片V_t和上一帧图片V_t－1做卷积,用V_t的卷积特征与V_t－1的卷积特征做减法,保留卷积特征差异性,再将卷积特征差与当前帧V_t的卷积特征相加,从而增强对V_t的空间特征卷积;双流卷积网络的时间卷积流对当前帧的光流图片V_t′进行时序特征卷积;最后将增强后的空间特征与时序特征融合进行分类。实验结果表明,改进后的双流卷积网络的识别率比原始的双流卷积网络提高了6.2%,且在公开数据集上的测试准确率达到了92.15%,从而证明了该方法的有效性和优越性。此外,与其他方法相比,该网络具有低深度、高识别率的特征,不仅能提高火灾和烟雾的识别率,而且实现了火灾的早期发现,缩短了检测时间。

关键词: 光流法, 开阔空间火灾监测, 空间特征增强网络, 时空特征融合, 双流卷积神经网络

Abstract: Fire detection based on image processing technology is an important branch in the field of fire monitoring in recent years.Aiming at the fire detection of open environment,using the dynamic and static characteristics of smoke and flame generated during the fire,the two-stream convolutional neural network is used as the theoretical basis to detect the fire.The two-stream convolutional neural network uses spatial and temporal streams to extract spatial information and temporal information in the video respectively.However,the information in the early stage of the fire is weak and the features are not obvious enough.In order to improve the initial recognition rate,a spatial enhancement network was proposed as the spatial stream of the two-stream convolutional neural network to extract and enhance the spatial information of the video.The spatial enhancement network simultaneously convolves the current frame V_t and the previous frame V_t－1,subtracting the convolution features of the V_t image with the convolution features of the V_t－1 image,to preserve the difference of the convolution features,and adding the convolution features difference to the convolution features of the current frame V_t,thereby enhance the spatial features convolution of the current frame V_t－1.Temporal stream of two-stream convolutional network convolves the optical flow image V_t′ of the current frame to get the temporal features.Finally,the enhanced spatial and temporal features are fused to classify.The experimental results show that the improved two-stream convolutional network has a 6.2% higher recognition rate than the original two-stream convolutional network,and achieved 92.15% recognition rate on the public dataset,indicating the effectiveness and superiority of the proposed method.Comparing with other methods,the network structure is designed lower but achieves good results,improves the identification accuracy of fire and smoke as well as realizes the early warning of fire,shorten detection time.

Key words: Open environment fire detection, Optical flow method, Spatial enhancement network, Spatial temporal features fusion, Two-stream convolutional network

中图分类号:

TP181

徐登, 黄晓东. 基于改进双流卷积网络的火灾图像特征提取方法[J]. 计算机科学, 2019, 46(11): 291-296. https://doi.org/10.11896/jsjkx.180901640

XU Deng, HUANG Xiao-dong. Fire Images Features Extraction Based on Improved Two-stream Convolution Network[J]. Computer Science, 2019, 46(11): 291-296. https://doi.org/10.11896/jsjkx.180901640

参考文献

[1]LI H Y.Research on large space fire monitoring technologybased on image processing [D].Chongqing:Chongqing University,2008.(in Chinese)
李洪玉.基于图像处理的大空间火灾监测技术的研究[D].重庆:重庆大学,2008.
[2]FERNANDEZ A,ALVAREZ M X,BIANCONI F.Texture description through histograms of equivalent patterns [J].Journal of Mathematical Imaging and Vision,2013,1(4):76-102.
[3]AHMED A.Machine Learning Approach for Adaptive Automated Smoke Region Detection [D].Swiss Federal Institute of Techology Zurich,2015:1735-1780.
[4]YANG J,CHEN F,ZHANG W.Visual-based smoke detection using support vector machine[C]∥2008 Fourth International Conference on Natural Computation.IEEE,2008:301-305.
[5]TUNG T X,KIM J M.An effective four-stage smoke-detectionalgorithm using video images for early fire-alarm systems [J].Fire Safety Journal,2011,46(5):276-282.
[6]ZHANG Q,XU J,XU L.Deep Convolutional Neural Networks for Forest Fire Detection[C]∥2016 International Forum on Management,Education and Information Technology Application.2016[7] XU Z G,XU J L.Automatic Fire Smoke Detection Based on Image Visual Features [C]∥International Conference on Computational Intelligence and Security Workshops (CISW 2007).2007:316-319[8]ZHAO L.Research on video-based fire smoke detection algorithm [D].Xiamen:Huaqiao University,2017 (in Chinese)
赵亮.基于视频的火灾烟雾检测算法的研究[D].厦门:华侨大学,2017.
[9]NODA S,UEDA K.Fire detection in tunnels using an image processing method[C]∥Proceeding of Vehicle Navigation and Information Systems Conference.1994:57-62.
[10]XU W S,TIAN C Z,FANG S M.Automatic fire identification based on image visual features [J].Computer Engineering,2003,29(18):112-113.(in Chinese)
许维胜,田长征,方盛明.基于图像视觉特征的火灾自动识别[J].计算机工程,2003,29(18):112-113.
[11]SONG W G,FAN W C,WU L B.Fire images detection method based on artificial neural network [J].Fire Science,1999(3):49-56.(in Chinese)
宋卫国,范维澄,吴龙标.基于人工神经网络的火灾图像探测方法[J].火灾科学,1999(3):49-56.
[12]SUN C.Research and Design of Fire Detection Algorithm Based on Video Image [D].Jinan:Shandong University,2018.(in Chinese)
孙琛.基于视频图像的火灾检测算法研究与设计[D].济南:山东大学,2018.
[13]TAO C,ZHANG J,WANG P.Smoke Detection Based on Deep Convolutional Neural Networks[C]∥International Conference on Industrial Informatics Computing Technology.2017:150-153.
[14]HORN B,SCHUNCK B.Determining optical flow[J].ArtificialIntelligence,1981,17:185-203.
[15]MEMIN E,PEREZ P.Hierarchical estimation and segmentation of dense motion fields [J].International Journal of Computer Vision,2002,46(2):129-155[16]ZHANG J H,ZHUANG J,DU H F,et al.One fire flame recognition algorithm based on video multi-feature fusion [J].Journal of Xi’an Jiaotong University,2006,40(7):811-814.(in Chinese)
张进华,庄健,杜海峰,等.一种基于视频多特征融合的火焰识别算法[J].西安交通大学报,2006,40(7):811-814.
[17]KOSIOREK A.神经网络中的注意力机制[J].机器人产业,2017(6):14-19.
[18]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolution neural networks[C]∥International Conference on Neural Information Processing Systems.Curran Associates Inc.,2012:1097-1105.
[19]SIMONYAN K,ZISSERMAN A.Two-Stream ConvolutionalNetworks for Action Recognition in Videos [J].Computational Linguistics,2014,1(4):568-576.
[20]JIA Y,SHELHAMER E,DONAHUE J,et al.Caffe:Convolutional architecture for fast feature embedding[C]∥Proceedings of the 22nd ACM international conference on Multimedia.ACM,2014:675-678.
[21]DONAHUE J,ANNE HENDRICKS L,GUADARRAMA S,et al.Long-Term Recurrent Convolutional Networks for Visual Recognition and Description[C]∥IEEE Conference on ComputerVision and Pattern Recognition (CVPR).2015:2625-2634.
[22]DENG J,DONG W,SOCHER R,et al.ImageNet:A large-scale hierarchical image database[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2009.
[23]SHARMA J,GRANMO O C,GOODWIN M,et al.Deep Convolutional Neural Networks for Fire Detection in Images[C]∥Engineering Applications of Neural Networks(EANN 2017).Cham:Springer,2017:183-193.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于改进双流卷积网络的火灾图像特征提取方法

Fire Images Features Extraction Based on Improved Two-stream Convolution Network

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 3

Metrics

本文评价

推荐阅读 0

[1]	胡燕花,唐鹏,金炜东,何正伟. 铁路视频序列的FOE的估计 Estimation of FOE for Railway Video Sequences 计算机科学, 2018, 45(7): 226-229. https://doi.org/10.11896／j.issn.1002-137X.2018.07.039
[2]	陈诚,常侃,莫彩网,李天亦,覃团发. 基于非局部均值和总变分最小化的单视频超分辨率算法 Single Video Super-resolution Algorithm Based on Non-local Means and Total Variation Minimization 计算机科学, 2018, 45(3): 263-267. https://doi.org/10.11896/j.issn.1002-137X.2018.03.042
[3]	刘赏,董林芳. 人群运动方向异常检测算法 Abnormal Crowd Movement Direction Detection Algorithm 计算机科学, 2013, 40(Z11): 337-340.