Abstract Forest fire is one of the most dangerous disasters that threaten the safety of human life and property. In order to detect fire in time, we detect the smoke when the fire breaks out. However, it is still a challenging task due to the variations of smoke in color, texture, shape and the disturbances of smoke-like objects. Therefore, the accuracy of smoke detection is not high, and it is accompanied by a high false positive rate, especially in the real environment. To tackle this problem, this paper proposes a novel model based on Faster Region-based Convolutional Network (R-CNN) which utilizes negative sample mining method. The proposed method allows the model to learn more negative sample features, thereby reducing false positives in smoke detection. The experiments are performed on self-created dataset containing 11958 images which are collected from cameras placed in villages or towns and existing datasets. Compared to other smoke datasets, the self-created dataset is larger and contains complex scenes. The proposed method achieves 94.59% accuracy, 94.35% precision and 5.76% false positive rate on self-created dataset. The results show that the proposed network is better and more robust than previous works.

References

[1] Abadi M., Agarwal A., Barham P., and et al., “TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems,” arXiv preprint arXiv:1603.04467, 2016.

[2] Cai Z. and Vasconcelos N., “Cascade R-CNN: Delving into High Quality Object Detection,” arxiv.org/abs/1712.00726v1, 2017.

[3] Cao Y., Yang F., Tang Q., and Lu X., “An Attention Enhanced Bidirectional LSTM for Early Forest Fire Smoke Recognition

[J],” IEEE Access, vol. 7, pp. 154732-154742, 2019.

[4] Chen T., Wu P., and Chiou Y., “An Early Fire- Detection Method Based on Image Processing,” in Proceedings of International Conference on Image Processing, Singapore, pp. 1707-1710, 2004.

[5] Chunyu Y., Jun F., Jinjun W., and Yongming Z., “Video Fire Smoke Detection Using Motion and Color Features,” Fire Technology, vol. 46, no. 3, pp. 651-663, 2010.

[6] Dai J., Li Y., and Sun J., “R-Fcn: Object Detection Via Region Based Fully Convolutional Networks,” Advances in Neural Information Processing Systems, vol. 29, 2016.

[7] Da-Penha O. and Nakamura E., “Fusing Light and Temperature Data for Fire Detection,” in Proceedings of IEEE Symposium on Computers and Communications, Riccione, pp. 107-112, 2010.

[8] Deng J., Guo J., Xue N., and Zafeiriou S., “Arcface: Additive Angular Margin Loss for Deep Face Recognition,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, pp. 4690- 4699, 2019.

[9] Everingham M., Van Gool L., Williams C., Winn J., and Zisserman A., “The Pascal Visual Object Classes (VOC) Challenge,” International Journal of Computer Vision, vol. 88, no. 2, pp. 303-338, 2010.

[10] Freund Y. and Schapire R., “A Decision- Theoretic Generalization of On-Line Learning and an Application Tto Boosting,” in Proceedings of European Conference on Computational Learning Theory, Barcelona, pp. 23-37, 1995.

[11] Ge Y., Zhang R., Wu L., Wang X., Tang X., and Luo P., “Deep Fashion2: A Versatile Benchmark for Detection, Pose Estimation, Segmentation and Re-Identification of Clothing Images,” arxiv.org/pdf/1901.07973.pdf, 2019.

[12] Haykin S., Kalman Filtering and Neural Networks, Wiley Online Library, 2001.

[13] He K., Gkioxari G., Dollár P., and Girshick R., “Mask r-cnn,” in Proceedings of the IEEE International Conference on Computer Vision, Honolulu, pp. 2961-2969, 2017.

[14] He K., Zhang X., Ren S., and Sun J., “Deep Residual Learning For Image Recognition,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 770-778, 2016.

[15] Horng W., Peng J., and Chen C., “A New Image-Based Real-Time Flame Detection Method Using Color Analysis,” in Proceedings of IEEE Networking, Sensing and Control, Tucson, pp. 100-105, 2005.

[16] Hu Y. and Lu X., “Real-Time Video fire Smoke Detection by Utilizing Spatial-Temporal Convnet Features,” Multimedia Tools and Applications, vol. 77, no. 8, pp. 29283-29301, 2018.

[17] Jiang B., Luo R., Mao J., Xiao T., and Jiang Y., “Acquisition of Localization Confidence for Accurate Object Detection,” in Proceedings of the European Conference on Computer Vision, pp. 784-799, 2018.

[18] Johnston F., Henderson S., Chen Y., Randerson J., Marlier M., DeFries R., Kinney P., Bowman DM., and Brauer M., “Estimated Global Mortality Attributable to Smoke From Landscape Fires,” Environmental Health Perspectives,” vol. 120, no. 5, pp. 695-701, 2012.

[19] Kaiser T., “Fire Detection with Temperature Sensor Arrays,” in Proceedings of IEEE 34th Annual International Carnahan Conference on Security Technology, Ottawa, pp. 262-268, 2000.

[20] Khan S., Muhammad K., Mumtaz S., Baik S., and De-Albuquerque V., “Instituto de Telecomunicações, Aveiro, Portugal Sejong University, Seoul, South Korea “Energy-Efficient Deep CNN for Smoke Detection in Foggy IoT Environment,” IEEE Internet of Things Journal, vol. 6, no. 6, pp. 9237-9245, 2019.

[21] Ko B., Kwak J., and Nam J., “Wildfire Smoke Detection Using Temporo-Spatial Features and Random Forest Classifiers,” Optical Engineering, vol. 51, no. 1, pp. 017208, 2012.

[22] Koman P., Billmire M., Baker K., De Majo R., Anderson F., Hoshiko S., Thelen B., and French 702 The International Arab Journal of Information Technology, Vol. 19, No. 4, July 2022 N., “Mapping Modeled Exposure of Wildland Fire Smoke for Human Health Studies in California,” Atmosphere, vol. 10, no. 6, pp. 308, 2019.

[23] Krizhevsky A., Sutskever I., and Hinton G., “ImageNet Classification with Deep Convolutional Neural Networks,” Advances in Neural Information Processing Systems, vol. 25, pp. 1097-1105, 2012.

[24] Liu W., Anguelov D., Erhan D., Szegedy C., Reed S., Fu C., and Berg A., “Ssd: Single Shot Multibox Detector,” in Proceedings of European Conference on Computer Vision, Amsterdam, pp. 21-37, 2016.

[25] Muhammad K., Ahmad J., and Baik S., “Early fire Detection Using Convolutional Neural Networks During Surveillance for Effective Disaster Management,” Neurocomputing, vol. 288, pp. 30-42, 2018.

[26] Neubeck A. and Van-Gool L., “Efficient Non- Maximum Suppression,” in Proceedings of 18th International Conference on Pattern Recognition, Hong Kong, pp. 850-855, 2006.

[27] Pang J., Chen K., Shi K., Feng H., Ouyang W., and Lin D., “Libra R-CNN: Towards Balanced Learning for Object Detectionl,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, pp. 821-830, 2019.

[28] Phillips W., Shah M., and Lobo N., “Flame Recognition in Video,” in Proceedings 5th IEEE Workshop on Applications of Computer Vision, Palm Springs, pp. 224-229, 2000.

[29] Pundir A. and Raman B., “Deep Belief Network for Smoke Detection,” Fire Technology, vol. 53, pp. 1943-1960, 2017.

[30] Redmon J., Divvala S., Girshick R., and Farhadi A., “You Only Look Once: Unified, Real-Time Object Detection,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 779-788, 2016.

[31] Ren S., He K., Girshick R., and Sun J., “Faster R- Cnn: Towards Real-Time Object Detection With Region Proposal Networks,” Advances in Neural Information Processing Systems, vol. 28, 2015.

[32] Renwei T., Zhongjie Z., Yongqiang B., Ming G., and Zhifeng G., “Key Parts Of Transmission Line Detection Using Improved Yolo V3,” The International Arab Journal Of Information Technology, vol. 18, no. 6, pp. 747-754, 2021.

[33] Simonyan K. and Zisserman A., “Very Deep Convolutional Networks for Large-Scale Image Recognition,” arXiv preprint arXiv:1409.1556, 2014.

[34] Sudharshan P., Petitjean C., Spanhol F., Oliveira L., Heutte L., and Honeine P., “Multiple Instances Learning for Histopathological Breast Cancer Image Classification,” Expert Systems with Applications, vol. 117, pp. 103-111, 2019.

[35] Toreyin B., Dedeoglu Y., and Cetin A., “Wavelet Based Real-Time Smoke Detection in Video,” in Processing of European Signal Processing Conference, Antalya, pp. 1-4, 2005.

[36] Toulouse T., Rossi L., Campana A., Celik T., and Akhloufi M., “Computer Vision for Wildfire Research: an Evolving Image Dataset for Processing and Analysis,” Fire Safety Journal, vol. 92, pp. 188-194, 2017.

[37] Vezzani R., Calderara S., Piccinini P., and Cucchiara R., “Smoke detection in video surveillance: The use of ViSOR (Video Surveillance On-line Repository,” in Proceedings of the International Conference on Content- based Image and Video Retrieval, Niagara Falls, pp. 289-298, 2008.

[38] Wen T., Edelman A., and Gorsich D., “A Fast Projected Conjugate Gradient Algorithm for Training Support Vector Machines,” Contemporary Mathematics, vol. 323, pp. 245- 263, 2001.

[39] Xu G., Zhang Y., Zhang Q., Lin G., Wang Z., Jia Y., and Wang J., “Video Smoke Detection Based On Deep Saliency Network, “Fire Safety Journal, vol. 105, pp. 277-285, 2019.

[40] Ye S., Bai Z., Chen H., Bohush R., and Ablameyko S., “An Effective Algorithm to Detect Both Smoke and Flame Using Color and Wavelet Analysis,” Pattern Recognition and Image Analysis, vol. 27, no. 1, pp. 131-138, 2017.

[41] Yin M., Lang C., Li Z., Feng S., and Wang T., “Recurrent Convolutional Network for Video- Based Smoke Detection,” Multimedia Tools Applications, vol. 78, no. 8, pp. 237-256, 2019.

[42] Yuan F., “A Fast Accumulative Motion Orientation Model Based on Integral Image for Video Smoke Detection,” Pattern Recognition Letters, vol. 29, no. 7, pp. 925-932, 2008.

[43] Zhang L., Zhu X., Chen X., Yang X., Lei Z., and Liu Z., “Weakly Aligned Cross-Modal Learning for Multispectral Pedestrian Detection,” in Processing of the IEEE International Conference on Computer Vision, Seoul, pp. 5127-5137, 2019.

[44] Zhang Q., Lin G., Zhang Y., and Xu G., and Wang J., “Wildland Forest Fire Smoke Detection Based on Faster R-CNN using Synthetic Smoke Images,” Procedia Engineering, vol. 211, pp. 441-446, 2018.

[45] Zhang Q., Xu J., Xu L., and Guo H., “Deep Convolutional Neural Networks for Forest fire Detection,” in Proceedings of International Forum on Management, Education and Information Technology Application, pp. 568- 575, 2016. Smoke Detection Algorithm based on Negative Sample Mining 703

[46] Zhao Y., Zhou Z., and Xu M., “Forest fire Smoke Video Detection Using Spatio-Temporal and Dynamic Texture Features,” Journal of Electrical and Computer Engineering, 2015.

[47] Zhao Y., “Candidate Smoke Region Segmentation of Fire Video Based on Rough Set Theory,” Journal of Electrical and Computer Engineering, 2015. Pei Ma comes from the School of Computer and Artifical Intelligence, Wuhan Textile University. Her major is computer science and technology, and her main research direction is computer vision and object detection. Fen g Yu is currently a lecturer with the school of computer science and intelligence, Wuhan Textile University. He received the Ph.D. degree with the school of computer science and technology, Huazhong University of Science and Technology. His research interests include machine vision algorithm, artificial intelligence application, and clothing intelligent manufacturing. Changlong Zhou is currently a experimentalist of Wuhan Textile University. His research interests include computer system architecture, artificial intelligence application, and computer application technology. Minghua Jiang is currently the vice-chancellor of Wuhan Textile University, and also is a professor with school of computer science and intelligence, Wuhan Textile University. He received the Ph. D. degree from the school of computer science and technology, Huazhong University of Science and Technology. His research interests include computer system architecture, artificial intelligence application, and clothing intelligent manufacturing