火災屬於意外或人為發生的災害，以致於有不少人員受傷甚至是死亡，在現實生活中有很多偵測火災發生時的設備，如偵煙探測器、火警自動警報設備、緊急廣播設備、住宅警報器、火警受信總機等等，即使如此偵測火災的發生仍然會有死角漏洞等缺陷，無法提早發出通報以減少人員生命財產的受損。
　　本論文使用YOLO (You Only Look Once)V4深度學習模型方法，此為一種影像物件辨識的方法，除了上述所列之設備外，也可利用影像辨識來偵測現場是否有火災，本文主要偵測的物件以都市的建築物和車輛為主，2種均為常見火災發生的案件，檢測建築物內外及火燒車輛的附近是否發生火災及人員在場，除了上述火災發生的預警通報外，也能協助消防隊判斷火災發生及人員的位置，以提升救援的效率。
實驗結果得知在適用於火災場景時，YOLOv4與YOLOv4 CSP具有較佳的檢測效果，YOLOv4的最佳mAP為85.67%，YOLOv4 CSP的最佳mAP為82%，YOLOv4 Tiny的最佳mAP僅69.27%，且有許多誤判及辨識的情況，因此除非在硬體條件比較嚴苛的情況下才考慮使用，否則理應以YOLOv4與YOLOv4 CSP為主要的辨識算法。

Fire is an accidental or human-caused disaster that results in injuries or even deaths for many people. There are many devices that detect fires in real life, such as smoke detectors, automatic fire alarm devices, emergency broadcast equipment, home alarms, fire alarm master stations, etc. However, even with such devices, there are still dead angles and other defects that prevent early notification to reduce the loss of lives and property.
　　This thesis uses the YOLO (You Only Look Once) V4 deep learning model method, which is an image object recognition method. In addition to the above-mentioned devices, image recognition can also be used to detect fires on-site. The main objects detected in this paper are urban buildings and vehicles, which are both common cases of fire.
　　The detection checks whether there is a fire and whether there are people inside and outside the buildings and near the burning vehicles, as well as assisting the fire department in determining the location of the fire and the people to improve the efficiency of rescue.
　　Experimental results show that YOLOv4 and YOLOv4 CSP have better detection results when applied to fire scenes. YOLOv4 has a best mAP of 85.67%, YOLOv4 CSP has a best mAP of 82%, and YOLOv4 Tiny has a best mAP of only 69.27%, with many misjudgments and recognition cases. Therefore, unless under hardware conditions that are more stringent, YOLOv4 and YOLOv4 CSP should be the main recognition algorithms, rather than YOLOv4 Tiny.

[1] 內政部消防署全球資訊網(2022)。火災統計。2022年9月15日，取自： https://www.nfa.gov.tw/cht/index.php?code=list&ids=220
[2] 林東清(2018)。資訊管理e化企業的核心競爭能力七版。臺北市：智勝文化。
[3] 李謦伊(2021)。YOLO演進—3—YOLOv4詳細介紹。2021年4月10日，取自：https://medium.com/ching-i/yolo演進-3-yolov4詳細介紹-5ab2490754ef
[4] Bochkovskiy, A., Wang, C.-Y., & Hong, Y. (2020). Yolov4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934.
[5] He, K., Zhang, X., Ren, S., & Sun, J. (2015). Spatial pyramid pooling in deep convolutional networks for visual recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 37(9), 1904-1916.
[6] He, K., Zhang, X., Ren, S., & Sun, J. (2015). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 770-778.
[7] Hinton, G. E., & Salakhutdinov, R. R. (2006). Reducing the dimensionality of data with neural networks. Science, 313(5786), 504-507.
[8] Hu, H., Wang, G., Zhang, Q., Wang, J., Fang, J., & Zhang, Y. (2009). Design wireless multi-sensor fire detection and alarm system based on ARM. In 2009 9th International Conference on Electronic Measurement & Instruments, pp. 3-285.
[9] LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
[10] Lee, Y., Hwang, J.-w., Lee, S., Bae, Y., & Park, J. (2019). An energy and GPU-computation efficient backbone network for real-time object detection. IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, 2019-June, 752–760. https://doi.org/10.1109/CVPRW.2019.00103
[11] Lin, T.-Y., Doll'ar, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017). Feature pyramid networks for object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2117-2125.
[12] Liu, S., Qi, L., Qin, H., Shi, J., & Jia, J. (2018). Path aggregation network for instance segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 8759-8768.
[13] Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C., & Berg, A. (2020). Ssd: Single shot multibox detector. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14, pp. 21-37. Springer International Publishing.
[14] Mohd Razmi, S., Saad, N., & Asirvadam, V. S. (2010). Vision-based flame detection: Motion detection & fire analysis. In 2010 IEEE Student Conference on Research and Development (SCOReD), pp. 187-191.
[15] Mwedzi, N. A., Nwulu, N. I., & Gbadamosi, S. L. (2019). Machine learning applications for fire detection in a residential building. In 2019 IEEE 6th International Conference on Engineering Technologies and Applied Sciences (ICETAS), pp. 1-4.
[16] Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 779-788.
[17] Redmon, J., & Farhadi, A. (2017). YOLO9000: better, faster, stronger. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 779-788.
[18] Redmon, J., & Farhadi, A. (2018). Yolov3: An incremental improvement. arXiv preprint arXiv:1804.02767.
[19] Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster r-cnn: Towards real-time object detection with region proposal networks. arXiv preprint arXiv:1506.01497.
[20] Rezatofighi, H., Tsoi, N., Gwak, J., Sadeghian, A., Reid, I., & Savarese, S. (2019). Generalized intersection over union: A metric and a loss for bounding box regression. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 658-666.
[21] Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
[22] Solorzano, A., Fonollosa, J., Fernandez, L., Eichmann, J., & Marco, S. (2017). Fire detection using a gas sensor array with sensor fusion algorithms. In 2017 ISOCS/IEEE International Symposium on Olfaction and Electronic nose (ISOEN), pp. 1-3.
[23] Soviany, P., & Ionescu, R. T. (2018). Optimizing the trade-off between single-stage and two-stage deep object detectors using image difficulty prediction. In 2018 20th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC), pp. 209-214.
[24] Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & Rabinovich, A. (2014). Going deeper with convolutions. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1-9.
[25] Wang, C.-Y., Bochkovskiy, A., & Liao, H.-Y. M. (2021). Scaled-yolov4: Scaling cross stage partial network. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 13029-13038.
[26] Wang, C.-Y., Hong, Y., Yeh, I. H., Wu, Y.-H., Chen, P.-Y., & Hsieh, J.-W. (2019). CSPNet: A new backbone that can enhance learning capability of CNN. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pp. 390-391.
[27] Wu, S., & Zhang, L. (2018). Using popular object detection methods for real time forest fire detection. In 2018 11th International Symposium on Computational Intelligence and Design (ISCID), pp. 280-284.
[28] Zheng, Z., Wang, P., Liu, W., Li, J., Ye, R., & Ren, D. (2020). Distance-IoU loss: Faster and better learning for bounding box regression. In Proceedings of the AAAI Conference on Artificial Intelligence, 34(7), 12993-13000.