CNN-Based Lightweight Flame Detection Method in Complex Scenes
LI Xinjian1,2, ZHANG Dasheng3, SUN Lilei4, XU Yong1,2
1. School of Computer Science and Technology, Harbin Institute of Technology(Shenzhen), Shenzhen 518055; 2. Shenzhen Key Laboratory of Visual Object Detection and Re-cognition, Harbin Institute of Technology(Shenzhen), Shen-zhen 518055; 3. Liangjiang Artificial Intelligence Academy, Chongqing University of Technology, Chongqing 401135; 4. College of Computer Science and Technology, Guizhou University, Guiyang 550025
Abstract：The existing fire detection methods rely on high-performance machines, and therefore the speeds on the embedded terminals and the mobile ones are not satisfactory. For most of the detection methods, the speed is low and the false detection rate is high, especially for small-scale fires missed detection problems. To solve these problems, a fire detection method based on you only look once is proposed. Depthwise separable convolution is employed to improve its network structure. Multiple data augmentation and bounding box based loss function are utilized to achieve a higher accuracy. The real-time 21ms fire detection on embedded mobile system is realized through parameter tuning with the detection accuracy ensured. Experimental results show that the proposed method improves accuracy and speed on the fire dataset.
[1] ZHOU X L,YU F X,WEN Y C,et al.Early Fire Detection Based on Flame Contours in Video.Information Technology Journal,2010,9(5):899-908. [2] SEEBAMRUNGSAT J,PRAISING S,RIYAMONGKOL P.Fire Detection in the Buildings Using Image Processing//Proc of the 3rd ICT International Student Project Conference.Washington,USA:IEEE,2014:95-98. [3] CHEN T H,WU P H,CHIOU Y C.An Early Fire-Detection Method Based on Image Processing//Proc of the International Conference on Image Processing.Washington,USA:IEEE,2004,III:1707-1710. [4] WANG T,SHI L,YUAN P,et al. A New Fire Detection Method Based on Flame Color Dispersion and Similarity in Consecutive Frames//Proc of the Chinese Automation Congress.Washington,USA:IEEE,2017:151-156. [5] 陈 磊,黄继风.基于视频的火焰检测方法.计算机工程与设计,2014,35(9):3143-3147,3195. (CHEN L,HUANG J F.Flame Detection Method Based on Video.Computer Engineering and Design,2014,35(9):3143-3147,3195.) [6] 王 浩,单文静,方宝富.基于多层上下文卷积神经网络的目标检测算法.模式识别与人工智能,2020,33(2):113-120. (WANG H,SHAN W J,FANG B F.Multi-layers Context Convolutional Neural Network for Object Detection.Pattern Recognition and Artificial Intelligence,2020,33(2):113－120.) [7] MUHAMMAD K,AHMAD J,MEHMOOD I,et al. Convolutional Neural Networks Based Fire Detection in Surveillance Videos.IEEE Access,2018,6:18174-18183. [8] 邓 军,姚涵文,王伟峰,等.基于优化InceptionV1的视频火焰超像素检测方法[J/OL].[2020-12-12].https://kns.cnki.net/kcms/detail/31.1690.TN.20200819.1059.012.html. (DENG J,YAO H W,WANG W F,et al. Detection Method for Video Flame Based on Optimized InceptionV1[J/OL].[2020-12-12].https://kns.cnki.net/kcms/detail/31.1690.TN.20200819.1059.012.html.) [9] 赵飞扬,罗 兵,林国军,等.基于改进 YOLOv3 的火焰检测.中国科技论文,2020,15(7):820-826. (ZHAO F Y,LUO B,LIN G J,et al.Flame Detection Based on Improved YOLOv3.China Sciencepaper,2020,15(7):820-826.) [10] REDMON J,DIVVALA S,GIRSHICK R,et al.You Only Look Once:Unified,Real-Time Object Detection//Proc of the IEEE Conference on Computer Vision and Pattern Recognition.Washington,USA:IEEE,2016:779-788. [11] REDMON J,FARHADI A.YOLO9000:Better,Faster,Stronger//Proc of the IEEE Conference on Computer Vision and Pattern Recognition.Washington,USA:IEEE,2017:6517-6525. [12] REDMON J,FARHADI A.YOLOv3:An Incremental Improvement[C/OL].[2020-12-12].https://arxiv.org/pdf/1804.02767v1.pdf. [13] BOCHKOVSKIY A,WANG C Y,LIAO H Y M.YOLOv4:Optimal Speed and Accuracy of Object Detection[C/OL].[2020-12-12].https://arxiv.org/pdf/2004.10934.pdf. [14] HOWARD A,SANDLER M,CHU G,et al. Searching for Mobi-lenetv3//Proc of the IEEE/CVF International Conference on Computer Vision.Washington,USA:IEEE,2019:1314-1324. [15] LIN T Y,MAIRE M,BELONGIE S,et al. Microsoft COCO:Common Objects in Context//Proc of the European Conference on Computer Vision.Berlin,Germany:Springer,2014:740-755. [16] ZHANG H Y,CISSE M,DAUPHIN Y N,et al. Mixup:Beyond Empirical Risk Minimization[C/OL].[2020-12-12].https://arxiv.org/pdf/1710.09412.pdf. [17] ZHENG Z H,WANG P,LIU W,et al.Distance-IoU Loss:Faster and Better Learning for Bounding Box Regression//Proc of the AAAI Conference on Artificial Intelligence.Palo Alto,USA:AAAI Press,2020:12993-13000. [18] 王延召,彭国华,延伟东.基于流形排序和联合连通性先验的显著性目标检测.模式识别与人工智能,2019,32(1):82-93. (WANG Y Z,PENG G H,YAN W D.Salient Object Detection Based on Manifold Ranking and Co-connectivity.Pattern Recognition and Artificial Intelligence,2019,32(1):82-93.) [19] TARG S,ALMEIDA D,LYMAN K.Resnet in Resnet:Generalizing Residual Architectures[C/OL].[2020-12-12].https://arxiv.org/pdf/1603.08029.pdf. [20] SRIVASTAVA N,HINTON G,KRIZHEVSKY A,et al.Dropout:A Simple Way to Prevent Neural Networks from Overfitting.The Journal of Machine Learning Research,2014,15(56):1929-1958. [21] HÜTTNER V,STEFFENS C R,DA COSTA BOTELHO S S.First Response Fire Combat:Deep Leaning Based Visible Fire Detection//Proc of the Latin American Robotics Symposium and Brazilian Symposium on Robotics.Washington,USA:IEEE,2017.DOI:10.1109/SBR-LARS-R.2017.8215312.
2021, Vol. 34 
