跳到主要內容

臺灣博碩士論文加值系統

(44.200.140.218) 您好!臺灣時間:2024/07/18 04:05
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:陳柏瑞
研究生(外文):CHEN, BO-RUEI
論文名稱:應用深度學習物件偵測技術於水上偵測與搜救之研究
論文名稱(外文):Applying Deep Learning Object Detection Techniques to Water Surface Object Detection and Rescue
指導教授:許巍嚴
指導教授(外文):HSU, WEI-YEN
口試委員:劉長萱許巍嚴蔡志鑫
口試委員(外文):LIOU, MICHELLEHSU, WEI-YENTSAI, CHIH-HSIN
口試日期:2022-07-12
學位類別:碩士
校院名稱:國立中正大學
系所名稱:資訊管理學系碩士在職專班
學門:電算機學門
學類:電算機一般學類
論文種類:學術論文
論文出版年:2022
畢業學年度:110
語文別:中文
論文頁數:75
中文關鍵詞:卷積神經網路物件偵測深度學習數據增強
外文關鍵詞:Convolutional Neural NetworkObject DetectionDeep LearningData Augmentation
相關次數:
  • 被引用被引用:1
  • 點閱點閱:150
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
隨著海上交通、海事工作者、水上遊憩的需求逐年增加,發生水域事故的機率也將隨之變高。如何在水上快速進行偵測並迅速展開搜索及救援是非常重要的任務,因此若能於高空進行拍攝並運用近年來迅速發展的影像辨識技術進行偵測,將可達到即時監測水上人員位置,並能夠在事故發生時快速知道水上人員確切位置並提高水上搜救存活的機率。
本研究以物件偵測技術 YOLO(You only look once)v4-Tiny 模型架構對空拍影像進行模型訓練。將探討針對不同數據增強方式以及調整模型超參數於水上物件偵測的可行性,主要做法為將數據進行前處理,再將影像進行數據增強,並調整不同的模型超參數,最終使用 mAP 指標與混淆矩陣來評估訓練結果。最終研究結果 mAP 最高達到 96.16% 和平均 Accuracy 為 95.85%。由研究結果可得知訓練後的 YOLO v4-Tiny 物件偵測模型可以有效的作為水上搜救之協助工具,更能夠守護每一位水上活動人員之安全。

As the demand for marine traffic, marine workers, and water recreation increases year by year, the probability of water accidents will also increase. How to quickly detect on the water and carry out search and rescue as soon as possible is a very important task. Therefore, if we can shoot at high altitude and use the image recognition technology that has developed rapidly in recent years for detection, we can achieve real-time monitoring of the position of personnel on the water. It is also able to quickly know the exact location of the person on the water when an accident occurs and increase the chance of survival in water search and rescue.
In this study, the object detection technique YOLO (You only look once) v4-Tiny model framework is used to train the model on aerial images. It will discuss the feasibility of different data enhancement methods and adjusting model hyperparameters for object detection on water. The main method is to pre-process the data, then enhance the image data, and adjust different model hyperparameters. Finally, the mAP metric and confusion matrix are used to evaluate the training results. Final study results mAP was up to 96.16% and Accuracy was 95.85%. From the study results, it can be known that the trained YOLO v4-Tiny object detection model can be used as an auxiliary tool for water search and rescue, and more effectively protect the safety of every water leisure personnel.

目錄 vi
圖目錄 viii
表目錄 x
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機 2
1.3 研究問題 5
1.4 研究目的 6
第二章 文獻探討 8
2.1 應用深度學習辨識水域事故之相關文獻 8
2.2 卷積神經網路 10
2.2.1 卷積層 Convolution Layer 11
2.2.2 線性整流單元 Rectified Linear Units Layer 12
2.2.3 池化層 Pooling Layer 13
2.2.4 全連接層 Fully Connected Layer 14
2.2.5 基本CNN架構及應用 15
2.3 優化神經網路 17
2.3.1 批次正規化 Batch Normalization 18
2.3.2 殘差學習 Residual Learning 20
2.4 YOLO 22
2.4.1 YOLO v1 23
2.4.2 YOLO v2 25
2.4.3 YOLO v3 27
2.4.4 YOLO v4 29
2.5 深度學習應用於物件偵測相關文獻 31
第三章 材料與研究方法 33
3.1 實驗材料 33
3.2 方法流程圖 35
3.3 資料前處理 35
3.4 數據增強 37
3.5 神經網路模型 39
3.5.1 Backbone 42
3.5.2 Neck 43
3.5.3 Head 44
第四章 實驗評估與結果 46
4.1 實驗環境 46
4.2 實驗評估指標 46
4.3 實驗結果 49
第五章 結論與未來展望 59
5.1 研究結論 59
5.2 研究限制 59
5.3 未來研究方向 60
參考文獻 61

中文文獻
內政部消防署(2020年04月)。救援案件之統計分析。消防月刊。https://ebook.nfa.gov.tw/10904/?fbclid=IwAR1wB8WONbNuaeyPLRxYVh5TQuuNzzjaBA7UAQ5OJMPRWw5nBCX1hbzH-10#p=32
中華民國行政院(2019年03月17日)。土地。行政院國情介紹。https://www.ey.gov.tw/state/4447F4A951A1EC45/094b1d53-de8d-4393-bde6-ab092969cce4
世界民報(2021年07月23日)。全球近三分之二的溺水死亡事件發生在亞太地區。World People News. http://www.worldpeoplenews.com/content/news/331223
陳允傑(2019年08月)。打造你的卷積神經網路。TensorFlow與Keras Python深度學習應用實務。
國立海洋科技博物館(2017年10月)。不一樣的環島 放眼東西南北四海岸。潮境專刊的5期。https://www.nmmst.gov.tw/chhtml/publication_detail/519/6/24
劉又銓(2020年12月02日)。衝浪運動常見傷害 預防大分析。康健。https://www.commonhealth.com.tw/blog/3887
Yang(2021年09月13日)。Yolov4相關演算法的原理及實現。大白智能第五章第四節。https://www.jiangdabai.com/video/%e5%8f%91%e5%b8%83%e8%a7%86%e9%a2%91%e6%b5%8b%e8%af%95-2-2-2-2-2-2-2-3
英文文獻
Bochkovskiy, A., Wang, C.-Y., & Liao, H.-Y. M. (2020). YOLOv4: Optimal Speed and Accuracy of Object Detection. http://arxiv.org/abs/2004.10934
Ghiasi, G., Lin, T. Y., & Le, Q.V. (2018). Dropblock: A regularization method for convolutional networks. Advances in Neural Information Processing Systems, 2018-Decem, 10727–10737.
Girshick, R. (2015). Fast R-CNN. Proceedings of the IEEE International Conference on Computer Vision, 2015 Inter, 1440–1448. https://doi.org/10.1109/ICCV.2015.169
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014). Rich feature hierarchies for accurate object detection and semantic segmentation. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 580–587. https://doi.org/10.1109/CVPR.2014.81
Hasan, S., Joy, J., Ahsan, F., Khambaty, H., Agarwal, M., & Mounsef, J. (2021). A Water Behavior Dataset for an Image-Based Drowning Solution. 1–5. https://doi.org/10.1109/igessc53124.2021.9618700
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2016-Decem, 770–778. https://doi.org/10.1109/CVPR.2016.90
Hirasawa, T., Aoyama, K., Tanimoto, T., Ishihara, S., Shichijo, S., Ozawa, T., Ohnishi, T., Fujishiro, M., Matsuo, K., Fujisaki, J., & Tada, T. (2018). Application of artificial intelligence using a convolutional neural network for detecting gastric cancer in endoscopic images. Gastric Cancer, 21(4), 653–660. https://doi.org/10.1007/s10120-018-0793-2
Huang, Z., Wang, J., Fu, X., Yu, T., Guo, Y., & Wang, R. (2020). DC-SPP-YOLO: Dense connection and spatial pyramid pooling based YOLO for object detection. Information Sciences, 522, 241–258. https://doi.org/10.1016/j.ins.2020.02.067
Ioffe, S., & Szegedy, C. (2015). Batch normalization: Accelerating deep network training by reducing internal covariate shift. 32nd International Conference on Machine Learning, ICML 2015, 1, 448–456.
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. Communications of the ACM, 60(6), 84–90. https://doi.org/10.1145/3065386
LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278–2323. https://doi.org/10.1109/5.726791
Lin, T.-Y., Dollár, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017). Feature Pyramid Networks for Object Detection. http://arxiv.org/abs/1612.03144
Liu, S., Qi, L., Qin, H., Shi, J., & Jia, J. (2018). Path Aggregation Network for Instance Segmentation. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 8759–8768. https://doi.org/10.1109/CVPR.2018.00913
Page, J., & Griffiths, T. (2014). Scanning and Surveillance: Swimming Pools, Beaches, Open Sea. Drowning, 323–329. https://doi.org/10.1007/978-3-642-04253-9_47
Quan, A., Herrmann, C., & Soliman, H. (2019). Project vulture: A prototype for using drones in search and rescue operations. Proceedings - 15th Annual International Conference on Distributed Computing in Sensor Systems, DCOSS 2019, 619–624. https://doi.org/10.1109/DCOSS.2019.00113
Ramírez, F. F., Benitez, D. S., Portas, E. B., & Orozco, J. A. L. (2011). Coordinated sea rescue system based on unmanned air vehicles and surface vessels. OCEANS 2011 IEEE - Spain. https://doi.org/10.1109/Oceans-Spain.2011.6003509
Rawat, W., & Wang, Z. (2017). Deep Convolutional Neural Networks for Image Classification: A Comprehensive Review. Neural Computation, 29(9), 2352–2449. https://doi.org/10.1162/neco_a_00990
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2016-Decem, 779–788. https://doi.org/10.1109/CVPR.2016.91
Redmon, J., & Farhadi, A. (2016). YOLO9000: Better, faster, stronger. Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, 2017-Janua, 6517–6525. https://doi.org/10.1109/CVPR.2017.690
Redmon, J., & Farhadi, A. (2018). YOLOv3: An Incremental Improvement. http://arxiv.org/abs/1804.02767
Ren, S., He, K., Girshick, R., & Sun, J. (2016). Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137–1149. https://doi.org/10.1109/TPAMI.2016.2577031
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. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2019-June, 658–666. https://doi.org/10.1109/CVPR.2019.00075
Simonyan, K., & Zisserman, A. (2015). Very deep convolutional networks for large-scale image recognition. 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings, 1–14.
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015). Going deeper with convolutions. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 07-12-June, 1–9. https://doi.org/10.1109/CVPR.2015.7298594
Wang, S., Han, Y., Chen, J., Zhang, Z., Wang, G., & Du, N. (2019). A Deep-Learning-Based Sea Search and Rescue Algorithm by UAV Remote Sensing. 2019 IEEE CSAA Guidance, Navigation and Control Conference (CGNCC), 1–5. https://doi.org/10.1109/GNCC42960.2018.9019134
World Health Organization. (2014, November 17). Global report on drowning: preventing a leading killer. https://www.who.int/publications/i/item/global-report-on-drowning-preventing-a-leading-killer
WHO (2021, April 27). Drowning. https://www.who.int/news-room/fact-sheets/detail/drowning
Yun, S., Han, D., Chun, S., Oh, S. J., Choe, J., & Yoo, Y. (2019). CutMix: Regularization strategy to train strong classifiers with localizable features. Proceedings of the IEEE International Conference on Computer Vision, 2019-Octob, 6022–6031. https://doi.org/10.1109/ICCV.2019.00612
Zheng, Z., Wang, P., Liu, W., Li, J., Ye, R., & Ren, D. (2020). Distance-IoU loss: Faster and better learning for bounding box regression. AAAI 2020 - 34th AAAI Conference on Artificial Intelligence, 2, 12993–13000. https://doi.org/10.1609/aaai.v34i07.6999
電子全文 電子全文(網際網路公開日期:20270724)
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
無相關期刊