請用此 Handle URI 來引用此文件：
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/96306完整後設資料紀錄
| DC 欄位 | 值 | 語言 |
|---|---|---|
| dc.contributor.advisor | 韓仁毓 | zh_TW |
| dc.contributor.advisor | Jen-Yu Han | en |
| dc.contributor.author | 黃晏辰 | zh_TW |
| dc.contributor.author | Yan-Chen Huang | en |
| dc.date.accessioned | 2024-12-24T16:16:08Z | - |
| dc.date.available | 2024-12-25 | - |
| dc.date.copyright | 2024-12-24 | - |
| dc.date.issued | 2024 | - |
| dc.date.submitted | 2024-11-15 | - |
| dc.identifier.citation | K. Amita. Hands-On Artificial Intelligence for IoT: Expert techniques for developing smarter IoT systems through Machine Learning and Deep Learning with Python. Packt Publishing, Birmingham, 2019. ISBN 1788836065.
W. Liang, J. Li, W. Diao, X. Sun, K. Fu, and Y. Wu. Fgatr-net: Automatic network architecture design for fine-grained aircraft type recognition in remote sensing images. Remote Sensing, 12(24):1–17, 2020. URL https://www.scopus.com/inward/record.uri?eid=2-s2.0-85098518546&doi=10.3390%2frs12244187&partnerID=40&md5=34e4d4baeed85da55bc39da7fe54fd5c. Cited By :5 Export Date: 8 September 2023 Article. G. Liu, X. Sun, K. Fu, and H. Q. Wang. Aircraft recognition in high-resolution satellite images using coarse-to-fine shape prior. IEEE GEOSCIENCE AND REMOTESENSING LETTERS, 10(3):573–577, 2013. ISSN 1545-598X 1558-0571 J9 - IEEE GEOSCI REMOTE S. doi: 10.1109/LGRS.2012.2214022. Times Cited in Web of Science Core Collection: 79 Total Times Cited: 91 Cited Reference Count: 17. Y. Nie, C. Bian, and L. Li. Adap-emd: Adaptive emd for aircraft fine-grained classification in remote sensing. IEEE Geoscience and Remote Sensing Letters, 19, 2022. URL https://www.scopus.com/inward/record.uri?eid=2-s2.0-85128638845&doi=10.1109%2fLGRS.2022.3168581&partnerID=40&md5=88af8bda7e7b7232270c71dede23eaca. Cited By :2 Export Date: 8 September 2023 Article. Y. Yi, Y. You, W. Zhou, and G. Meng. Mha-cnn: Aircraft fine-grained recognition of remote sensing image based on multiple hierarchies attention. In International Geoscience and Remote Sensing Symposium (IGARSS), volume 2022-July, pages443051–3054. URL https://www.scopus.com/inward/record.uri?eid=2-s2.0-85140374925&doi=10.1109%2fIGARSS46834.2022.9884219&partnerID=40&md5=88632c6d6c555005999476f4be098977. Cited By :2 Export Date: 8 September2023 Conference Paper. D. Cai. Study on aircraft recognition in high spatial resolution remote sensing image based on skeleton characteristics analysis. Advanced Materials Research, 268-270: 1982–1985, 2011. ISSN 1662-8985. doi: 10.4028/www.scientific.net/AMR.268-270. 1982. URL https://www.scientific.net/AMR.268-270.1982. Q. Wu, H. Sun, X. Sun, D. Zhang, K. Fu, and H. Wang. Aircraft recognition in high-resolution optical satellite remote sensing images. IEEE Geoscience and Remote Sensing Letters, 12(1):112–116, 2015. ISSN 1558-0571. doi: 10.1109/LGRS.2014.2328358. A. Zhao, K. Fu, S. Wang, J. Zuo, Y. Zhang, Y. Hu, and H. Wang. Aircraft recognition based on landmark detection in remote sensing images. IEEE Geoscience and Remote Sensing Letters, 14(8):1413–1417, 2017. ISSN 1558-0571. doi: 10.1109/LGRS.2017.2715858. Ben YOUSSEF Youssef, Merrouchi Mohamed, Abdelmounim Elhassane, and Gadi Taoufiq. Classification of aircraft in remote sensing images based on deep convolutional neural networks. Statistics, Optimization amp; Information Computing, 10(1), 2022. doi: 10.19139/soic-2310-5070-1143. URL http://www.iapress.org/index.php/ soic/article/view/1143. M. R. Mohammadi. Deep multiple instance learning for airplane detection in high-resolution imagery. MACHINE VISION AND APPLICATIONS, 32(1), 2021. ISSN 0932-8092 1432-1769 J9 - MACH VISION APPL. doi: 10.1007/s00138-020-01153-7. Times Cited in Web of Science Core Collection: 4 Total Times Cited: 4 Cited Reference Count: 71. Z. Z. Wu, S. H. Wan, X. F. Wang, M. Tan, L. Zou, X. L. Li, and Y. Chen. A benchmark data set for aircraft type recognition from remote sensing images. APPLIED SOFTCOMPUTING, 89, 2020. ISSN 1568-4946 1872-9681 J9 - APPL SOFT COMPUT.doi: 10.1016/j.asoc.2020.106132. Times Cited in Web of Science Core Collection: 33 Total Times Cited: 35 Cited Reference Count: 54 K. Gao, H. He, D. Lu, L. Xu, L. Ma, and J. Li. Optimizing and evaluating swintransformer for aircraft classification: Analysis and generalizability of the mtarsi dataset. IEEE Access, 10:134427–134439, 2022. URL https://www.scopus.com/inward/record.uri?eid=2-s2.0-85146226496&doi=10.1109%2fACCESS. 2022.3231327&partnerID=40&md5=e204469400b62bd9a1b4fe9701e611d5. Export Date: 8 September 2023 Article. Y. Yan, Y. Zhang, and N. Su. A novel data augmentation method for detection of specific aircraft in remote sensing rgb images. IEEE Access, 7:56051–56061, 2019. ISSN 2169-3536. doi: 10.1109/ACCESS.2019.2913191. J. Shermeyer, T. Hossler, A. V. Etten, D. Hogan, R. Lewis, and D. Kim. Rareplanes: Synthetic data takes flight. In 2021 IEEE Winter Conference on Applications of ComputerVision (WACV), pages 207–217, 2021. ISBN 2642-9381. doi: 10.1109/WACV48630.2021.00025. X. Leng, K. Ji, and G. Kuang. Radio frequency interference detection and localization in sentinel-1 images. IEEE Transactions on Geoscience and Remote Sensing, 59(11): 9270–9281, 2021. ISSN 1558-0644. doi: 10.1109/TGRS.2021.3049472. M. Z. Wong, K. Kunii, M. Baylis, W. H. Ong, P. Kroupa, and S. Koller. Synthetic dataset generation for object-to-model deep learning in industrial applications. PeerJ ComputSci, 5:e222, 2019. ISSN 2376-5992 (Electronic) 2376-5992 (Linking). Wong, Matthew Z Kunii, Kiyohito Baylis, Max Ong, Wai Hong Kroupa, Pavel Koller, Swen United States PeerJ. Computer science. B. Huang E. Gu W. Guo B. He, X. Li and L. Wu. Unityship: A large-scale synthetic dataset for ship recognition in aerial images, 2021. W. Armstrong, S. Drakontaidis, and N. Lui. Synthetic data for semantic image segmentation of imagery of unmanned spacecraft. In 2023 IEEE Aerospace Conference, pages 1–7, 2022. ISBN 1095-323X. doi: 10.1109/AERO55745.2023.10115564. B. Kiefer, D. Ott, and A. Zell. Leveraging synthetic data in object detection on unmanned aerial vehicles. In 2022 26th International Conference on Pattern Recognition (ICPR), pages 3564–3571, 2022. doi: 10.1109/ICPR56361.2022.9956710. URL http://doi.ieeecomputersociety.org/10.1109/ICPR56361.2022.9956710.46 J. Tremblay, A. Prakash, D. Acuna, M. Brophy, V. Jampani, C. Anil, T. To, E. Cameracci, S. Boochoon, and S. Birchfield. Training deep networks with synthetic data: Bridging the reality gap by domain randomization. In 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pages 1082–10828, 2018. doi:10.1109/CVPRW.2018.00143. URL http://doi.ieeecomputersociety.org/10.1109/CVPRW.2018.00143. Ma Yuchi, Chen Shuo, Ermon Stefano, and Lobell David B. Transfer learning in environmental remote sensing. Remote Sensing of Environment, 301:113924, 2024. ISSN 0034-4257. doi: https://doi.org/10.1016/j.rse.2023.113924. URL https://www.sciencedirect.com/science/article/pii/S0034425723004765. Tobin Josh, Fong Rachel, Ray Alex, Schneider Jonas, Zaremba Wojciech, and Abbeel Pieter. Domain randomization for transferring deep neural networks from simulation to the real world. In 2017 IEEE/RSJ international conference on intelligent robots and systems (IROS), pages 23–30. IEEE, 2017. ISBN 1538626829. S. I. Nikolenko. Directions in Synthetic Data Development, pages 227–234. Springer International Publishing, Cham, 2021. ISBN 978-3-030-75178-4. doi: 10.1007/978-3-030-75178-4_9. URL https://doi.org/10.1007/978-3-030-75178-4_9. Schieber Hannah, Demir Kubilay Can, Kleinbeck Constantin, Yang Seung Hee, and Roth Daniel. Indoor synthetic data generation: A systematic review. Computer Vision and Image Understanding, 240:103907, 2024. ISSN 1077-3142. doi: https://doi.org/10.1016/j.cviu.2023.103907. URL https://www.sciencedirect.com/science/article/pii/S1077314223002874. Zhu Jun-Yan, Park Taesung, Isola Phillip, and A. Efros Alexei. Unpaired image-to-image translation using cycle-consistent adversarial networks. CoRR, abs/1703.10593, 2017.URL http://arxiv.org/abs/1703.10593. Maji Subhransu, Rahtu Esa, Kannala Juho, Blaschko Matthew, and Vedaldi Andrea. Finegrained visual classification of aircraft. CoRR, abs/1306.5151, 2013. URL http://arxiv.org/abs/1306.5151. He Kaiming, Zhang Xiangyu, Ren Shaoqing, and Sun Jian. Deep residual learning for 47 image recognition. CoRR, abs/1512.03385, 2015. URL http://arxiv.org/abs/1512.03385. Pan Sinno and Yang Qiang. A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering, 22(10):1345–1359, 2010. ISSN 1558-2191. doi:10.1109/TKDE.2009.191. M.L. McHugh. Interrater reliability: the kappa statistic. Biochemia medica, 22(3):276–282, 2012. ISSN 1330-0962. Desai Saurabh and G Ramaswamy Harish. Ablation-cam: Visual explanations for deep convolutional network via gradient-free localization. In 2020 IEEE Winter Conference on Applications of Computer Vision (WACV), pages 972–980, 2020. doi: 10.1109/ WACV45572.2020.9093360. | - |
| dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/96306 | - |
| dc.description.abstract | 衛星影像飛機型號識別在軍事與商業情報收集中扮演著重要角色，但高品質資料的收集卻面臨許多挑戰。本研究使用 Blender 3D 渲染軟體依照領域隨機化與領域自適應等理論，製作四種不同類型的合成資料。並將這些合成資料使用不同的資料量先行預訓練再使用真實影像參數微調，使用 ResNet-50 深度學習模型進行訓練。本研的目標在究探討合成資料的在飛機型號識別的表現與限制，並進一步了解不同合成資料製作方式與資料使用量對後續訓練結果的影響。研究結果顯示, 合成資料具有一定的效果。單獨使用合成資料進行訓練即可達到 60~70% 的準確率，結合真實資料進行參數微調後，準確率可進一步提升至 90% 以上。相較於僅使用真實資料的基線模型，此方法能提高 3~5% 的準確率。另一方面，本研究也發現合成資料使用存在一定限制。四種不同類型的合成資料在最終準確率上並無顯著差異，代表合成資料的處理方式可能不如預期重要。隨著合成資料量增加，準確率提升效果逐漸減弱，顯示單純增加合成資料量並不能持續提高模型的性能，也反映著真實資料的必要性。 | zh_TW |
| dc.description.abstract | Aircraft model recognition using satellite imagery is essential in military and commercial intelligence but faces challenges in acquiring high-quality data. This study uses Blender 3D to create four types of synthetic data, applying domain randomization and daptation theories. These datasets were used for pre-training at different volumes, followed by fine-tuning on real images with the ResNet-50 model. The aim is to assess synthetic data＇s effectiveness and limitations in aircraft recognition.
Results show that synthetic data alone can achieve 60 ∼ 70% accuracy, and finetuning with real data boosts this to over 90%, outperforming a real-data-only baseline by 3 ∼ 5%. However, increasing synthetic data volume has diminishing returns, highlighting the essential role of real data. | en |
| dc.description.provenance | Submitted by admin ntu (admin@lib.ntu.edu.tw) on 2024-12-24T16:16:07Z No. of bitstreams: 0 | en |
| dc.description.provenance | Made available in DSpace on 2024-12-24T16:16:08Z (GMT). No. of bitstreams: 0 | en |
| dc.description.tableofcontents | 口試委員審定書 i
摘要 iii Abstract iv 目次 v 圖次 vii 表次 viii 第一章 前言 1 1.1 研究背景 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 研究動機與目的 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 研究流程 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.4 論文架構 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 第二章 文獻回顧 4 2.1 衛星影像航空器型號識別相關研究 . . . . . . . . . . . . . . . . . . 4 2.2 現有飛機型號影像識別資料集盤點 . . . . . . . . . . . . . . . . . . 5 2.2.1 衛星飛機影像相關資料集 . . . . . . . . . . . . . . . . . . . . . . 5 2.2.2 合成資料應用於飛機辨識 . . . . . . . . . . . . . . . . . . . . . . 7 2.3 合成資料用於深度學習 . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.3.1 應用合成資料的深度學習研究 . . . . . . . . . . . . . . . . . . . 8 2.3.2 合成資料的特色與優勢 . . . . . . . . . . . . . . . . . . . . . . . 9 2.4 資料域與領域偏移 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.5 領域自適應 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.6 領域隨機化 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.7 小結 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 第三章 研究方法 12 3.1 合成資料集製作 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.1.1 合成資料集設計與環境參數設定 . . . . . . . . . . . . . . . . . . 13 3.1.2 合成環境光線與相機參數範圍設定 . . . . . . . . . . . . . . . . 15 3.1.3 以 CycleGAN 進行合成資料風格轉換 . . . . . . . . . . . . . . . 16 3.1.4 合成資料製作類型 . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.2 真實飛機衛星影像資料收集 . . . . . . . . . . . . . . . . . . . . . . 21 3.2.1 飛機型號選擇 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.2.2 真實飛機衛星影像資料收集數量 . . . . . . . . . . . . . . . . . . 22 3.3 深度學習模型與訓練 . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.3.1 深度學習模型 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.3.2 移轉學習影像風格轉換 . . . . . . . . . . . . . . . . . . . . . . . 23 3.3.3 深度學習實驗設計 . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.3.4 深度學習模型訓練超參數設定 . . . . . . . . . . . . . . . . . . . 25 3.4 影像分類評估參數與方法 . . . . . . . . . . . . . . . . . . . . . . . . 26 3.5 小結 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 第四章 實驗與結果分析 28 4.1 準確率與資料使用量間的關係 . . . . . . . . . . . . . . . . . . . . . 28 4.1.1 只使用合成資料時的準確率 . . . . . . . . . . . . . . . . . . . . 34 4.1.2 與只使用真實資料的模型比較 . . . . . . . . . . . . . . . . . . . 35 4.1.3 四種合成資料處理方式的準確率差異 . . . . . . . . . . . . . . . 37 4.2 各類飛機型號的準確率 . . . . . . . . . . . . . . . . . . . . . . . . . 38 4.3 小結 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 第五章 結論與建議 42 5.1 結論 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 5.2 建議與未來工作 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 參考文獻 44 | - |
| dc.language.iso | zh_TW | - |
| dc.subject | 影像分類 | zh_TW |
| dc.subject | 合成資料 | zh_TW |
| dc.subject | 飛機型號辨識 | zh_TW |
| dc.subject | Synthetic Data | en |
| dc.subject | Aircraft Type Recognition | en |
| dc.subject | Image Classification | en |
| dc.title | 以渲染合成資料於飛機型號識別模型訓練 | zh_TW |
| dc.title | Rendered Synthetic Data in Aircraft Type Recognition | en |
| dc.type | Thesis | - |
| dc.date.schoolyear | 113-1 | - |
| dc.description.degree | 碩士 | - |
| dc.contributor.oralexamcommittee | 吳日騰;甯方璽;曾子榜 | zh_TW |
| dc.contributor.oralexamcommittee | Rih-Teng Wu;Fang-Shii Ning;Tzu-Pang Tseng | en |
| dc.subject.keyword | 影像分類,合成資料,飛機型號辨識, | zh_TW |
| dc.subject.keyword | Image Classification,Synthetic Data,Aircraft Type Recognition, | en |
| dc.relation.page | 48 | - |
| dc.identifier.doi | 10.6342/NTU202404541 | - |
| dc.rights.note | 同意授權(限校園內公開) | - |
| dc.date.accepted | 2024-11-15 | - |
| dc.contributor.author-college | 工學院 | - |
| dc.contributor.author-dept | 土木工程學系 | - |
| dc.date.embargo-lift | 2027-12-31 | - |
| 顯示於系所單位： | 土木工程學系 | |
文件中的檔案：
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| ntu-113-1.pdf 未授權公開取用 | 12.54 MB | Adobe PDF | 檢視/開啟 |
系統中的文件，除了特別指名其著作權條款之外，均受到著作權保護，並且保留所有的權利。