跳到主要內容

臺灣博碩士論文加值系統

(44.213.60.33) 您好!臺灣時間:2024/07/20 05:32
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:魏呈宇
研究生(外文):Wei, Cheng-Yu
論文名稱:整合深度學習及3D Web GIS技術建立三維建物價值估算應用平台
論文名稱(外文):Integrating Deep Learning and 3D Web GIS Technology to Establish 3D Building Appraisal System
指導教授:饒見有饒見有引用關係
指導教授(外文):Rau, Jiann-Yeou
口試委員:趙鍵哲林昭宏
口試委員(外文):Jaw, Jen-JerLin, Chao-Hung
口試日期:2023-07-11
學位類別:碩士
校院名稱:國立成功大學
系所名稱:測量及空間資訊學系
學門:工程學門
學類:測量工程學類
論文種類:學術論文
論文出版年:2023
畢業學年度:111
語文別:中文
論文頁數:93
中文關鍵詞:深度學習語義分割屋頂材質分類三維地理資訊平台
外文關鍵詞:Deep learningSemantic segmentationRoof materials classification3D Web GIS
相關次數:
  • 被引用被引用:0
  • 點閱點閱:112
  • 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
近年來,軟硬體設備效能提升、自動化需求提升以及多元測繪技術進步,逐漸改變了傳統測量工程作業模式,無人機輕巧與不受地形限制之特性,再加上能夠事先規劃飛行路徑、遠端控制大幅減少了作業時間成本,都使其成為現今政府與私人機關進行測量工程規劃使用之工具。無人機影像之高解析度與多元應用成為其優勢,透過飛行航高與相機焦距設定,決定適合目標使用之地面解析度,除此之外,無人機影像之後續加值利用也為其一大價值,透過高重疊率與傾斜攝影使其能夠進行真實正射影像建置,真實正射影像除了消除高差移位外,其也為含有地理資訊之影像,對於往後需要進行地理資訊結合時提供座標基準,以相同座標基準將多元資料進行整合,因此以無人機影像進行影像辨識成為熱門議題。
影像辨識方法從發展至今也大不相同,由傳統監督式分類至當今熱門之深度學習,深度學習模型透過大量訓練資料進行學習,而能夠對不同實驗區之資料進行現象預測,深度學習模型中,大方向可以將其分為物件偵測與語義分割,其中物件偵測模型於偵測成果中,常會以物件框在最外圍對物件進行標記,此情形會造成辨識成果並不是貼合物件邊界,而語義分割模型會以每個像元為單位進行偵測,因此其分類之物件邊界會較貼近真實物件邊界,因此於模型訓練時能夠選擇符合需求的模型作為分類工具。
世界各地之建築物風格大不相同,台灣常見之鐵皮屋、磚瓦屋都成為一大特色,多雨潮濕之氣候特徵,造就與其他國家不同之建築模式,為排水與便利而設計之鐵皮斜屋頂、為散熱而建造之磚瓦屋頂,最後為空間利用而建築之水泥平屋頂,都為台灣常見屋頂材質形式,然而,位於亞熱帶季風氣候的台灣,豪雨與颱風侵襲過後,便會造成建築物與屋頂破損,除了評估一地何種屋頂材質易因氣候災害而損壞,提供適合之屋頂材質建議外,房屋價值受到許多因素影響,如政治、交通等常見之因素,屋頂材質對於房屋價值之影響或許不如其他條件有影響力,然而,不同屋頂材質將有不同重置成本,造就房屋價值不同,因此,初步以自動化判釋屋頂材質將房屋價值估計進行連結,嘗試將社會現象視覺化呈現。
三維地理資訊為近年新興發展之課題,三維地理資訊展示方式相較二維地理資訊更為貼近日常生活,因此智慧城市、數位孿生等三維地理資訊概念逐漸蓬勃,以三維平台進行現象展示日趨重要,使用最廣為通用之CityGML資料格式進行視覺化也為當前許多國政府致力進行之工作內容,因此,如何將研究議題以三維地理資訊平台展示成為熱門主題。
基於以上科技發展趨勢與台灣環境需求條件下,本研究將涵蓋資料前處理至視覺化展示,先以無人機進行影像蒐集,將其進行高差移位修正以利往後資料整合應用,並且測試不同以語義分割深度學習模型進行屋頂材質影像辨識,最終,將屋頂材質辨識成果、房屋價值估計以三維地理資訊平台進行展示。
Recently, the improvement of software and the hardware equipment, the increasing demand of automation requirements, and the advancement of multiple surveying and mapping technologies have changed the traditional way of survey. UAV is a robot that is lightweight and unrestricted by the terrain. Otherwise, we can plan the flight path in advance and remote control the flight which significantly reduces time consumption and is labor-intensive. As a result, UAV has become a useful tool for government and company to plan their work. UAV images also have the advantages of high resolution and multiple applications. Because of the high percentages of coverage in UAV images, they can be used to generate the true ortho image which removes the height relief problem and contains geometry information as coordinates. The coordinate information can be a reference to help the integration of multiple data for the applications. Therefore, using UAV images for image recognition has become a hot topic.
There are several methods in image classification and recognition, from traditional supervised methods to deep learning methods. Deep learning methods learn features from a mount of training data, and then the model can classify the features in another study area. Deep learning methods can also be divided into object detection and semantic segmentation. The classification results from object detection methods will be surrounded by the sliding window which does not match the object boundary. On the other hand, the semantic segmentation method detects all pixels and gives them the class. The result from this method will have a similar boundary as the origin object. Consequently, we can choose the classification method which matches our target.
The rainy and humid climate makes a different architectural appearance in Taiwan. The common roof materials are metal sheets which are usually attached to the sloping roof for the drain purpose, the tiles are built for heat dissipation, and the cement is usually in the flat roof for space utilization. However, because of the subtropical climate, Taiwan usually has lots of building and roof damage after heavy rain and typhoon. Additionally, different roof materials will have different replacement prices, used years, and so on. That is, how to identify the roof materials and asses their value is important.
3D geographical information has received great attention from different fields. The way of displaying 3D geographic information is closer to daily life than 2D geographic information. Therefore, the concept of smart city, digital twin, and so on have become flourishing. Many governments have worked on using common data formats to visualize the social phenomenon on the 3D platform. Based on the above technological development trends and Taiwan’s environmental requirements, this research will cover data preprocessing, image recognition to visualization display.
摘要 I
致謝 IX
目錄 XI
表目錄 XIV
圖目錄 XV
壹. 緒論 1
1.1 研究背景 1
1.2 研究目標 3
1.3 研究流程 4
1.3.1 文獻回顧 4
1.3.2 以深度學習模型進行屋頂材質分類 5
1.3.3 進行房屋價值估計 5
1.3.4 三維地理資訊平台呈現 6
1.3.5 結論與未來工作建議 6
1.4 論文架構 8
貳. 文獻回顧及探討 10
2.1 無人機影像取得及其加值應用 10
2.1.1 無人機影像介紹 10
2.1.2 無人機影像之應用 11
2.2 影像分類方法介紹 12
2.2.1 監督式分類 13
2.2.2 非監督式分類 13
2.2.3 人工智慧分類 14
2.3 精度分析及評估 23
2.4 三維建築物模型及應用 24
2.4.1 三維建築物模型介紹 24
2.4.2 自動化偵測屋頂應用 26
2.4.3 三維建築物模型及房價估計 27
參. 研究方法與步驟 29
3.1 研究流程 30
3.1.1 資料前處理 30
3.1.2 屋頂材質偵測 31
3.1.3 屋頂邊界篩選 32
3.1.4 精度分析 32
3.1.5 三維房屋模型建置 33
3.1.6 視覺化呈現 33
3.2 實驗資料簡介 34
3.2.1 實驗區介紹 34
3.2.2 使用資料介紹 35
3.3 研究工具介紹 36
3.3.1 Agisoft Metashape 36
3.3.2 ArcGIS Pro 36
3.3.3 Cesium ion 36
3.4 分類方法介紹 37
3.4.1 FCN 37
3.4.2 U-Net 38
3.4.3 DeepLabv3+ 39
3.5 精度分析 40
3.6 三維建物模型模擬展示及房屋估價 41
3.6.1 三維建物模型建立 41
3.6.2 房屋估價 43
肆. 研究成果 46
4.1 資料前處理 46
4.1.1 真實正射影像製作 49
4.1.2 訓練資料及地真資料製作 51
4.2 屋頂材質預測 51
4.2.1 DeepLabv3+於訓練屬性有無顏色條件之比較 51
4.2.2 FCN、U-Net 、DeepLabv3+於有顏色條件下之表現比較 58
4.2.3 小結 71
4.3 建置三維建築物模型 72
4.4 基於3D WEBGIS之視覺化呈現 78
4.4.1 屋頂材質之視覺化 79
4.4.2 房屋估價之視覺化 82
4.4.3 小結 84
伍. 結論及未來展望 85
5.1 結論 85
5.2 未來展望 87
參考文獻 88
Ball, G. H. & Hall, D. J. (1965). ISODATA, a novel method of data analysis and pattern classification. Stanford research inst Menlo Park CA.
Cai, Y., He, H., Yang, K., Fatholahi, S. N., Ma, L., Xu, L. & Li, J. (2021). A comparative study of deep learning approaches to rooftop detection in aerial images. Canadian Journal of Remote Sensing, 47(3), 413-431.
Carranza-García, M., Torres-Mateo, J., Lara-Benítez, P. & García-Gutiérrez, J. (2020). On the performance of one-stage and two-stage object detectors in autonomous vehicles using camera data. Remote Sensing, 13(1), 89.
Chen, L.-C., Papandreou, G., Kokkinos, I., Murphy, K. & Yuille, A. L. (2014). Semantic image segmentation with deep convolutional nets and fully connected crfs. arXiv preprint arXiv:1412.7062.
Chen, L.-C., Papandreou, G., Schroff, F. & Adam, H. (2017). Rethinking atrous convolution for semantic image segmentation. arXiv preprint arXiv:1706.05587.
Chen, L.-C., Zhu, Y., Papandreou, G., Schroff, F. & Adam, H. (2018). Encoder-decoder with atrous separable convolution for semantic image segmentation. Paper presented at the Proceedings of the European conference on computer vision (ECCV).
Chicco, D. & Jurman, G. (2020). The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation. BMC genomics, 21, 1-13.
Ciresan, D., Giusti, A., Gambardella, L. & Schmidhuber, J. (2012). Deep neural networks segment neuronal membranes in electron microscopy images. Advances in neural information processing systems, 25.
Döllner, J. & Buchholz, H. (2005). Continuous level-of-detail modeling of buildings in 3D city models. Paper presented at the Proceedings of the 13th annual ACM international workshop on Geographic information systems.
Du, L., Zhang, R. & Wang, X. (2020). Overview of two-stage object detection algorithms. Paper presented at the Journal of Physics: Conference Series.
Gröger, G., Kolbe, T. H., Nagel, C. & Häfele, K. H. (2012). OGC city geography markup language (CityGML) encoding standard.
Han, N., Zhang, W. & Liang, K. (2015). The design and implementation of Shenzhen house price indexes system based on 3D-GIS. Paper presented at the 2015 23rd International Conference on Geoinformatics.
He, K., Gkioxari, G., Dollár, P. & Girshick, R. (2017). Mask r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 2961-2969).
He, K., Zhang, X., Ren, S. & Sun, J. (2016). Deep residual learning for image recognition. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition.
Henn, A., Gröger, G., Stroh, V. & Plümer, L. (2013). Model driven reconstruction of roofs from sparse LIDAR point clouds. ISPRS Journal of photogrammetry and remote sensing, 76, 17-29.
Hu, J., You, S. & Neumann, U. (2003). Approaches to large-scale urban modeling. IEEE Computer Graphics and Applications, 23(6), 62-69.
Iradaf, M. & Rau, J. (2023). Reconstruction of LOD-2 3D building model using UAV imagery for building property application. Paper presented at the 第40屆測量及空間資訊研討會, 台灣台中市.
Janiesch, C., Zschech, P. & Heinrich, K. (2021). Machine learning and deep learning. Electronic Markets, 31(3), 685-695.
Krishna, K. & Murty, M. N. (1999). Genetic K-means algorithm. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 29(3), 433-439.
Kutzner, T., Chaturvedi, K. & Kolbe, T. H. (2020). CityGML 3.0: New functions open up new applications. PFG–Journal of Photogrammetry, Remote Sensing and Geoinformation Science, 88(1), 43-61.
Kim, J., Bae, H., Kang, H. & Lee, S. G. (2021). CNN algorithm for roof detection and material classification in satellite images. Electronics, 10(13), 1592.
Lewis, H. G. & Brown, M. (2001). A generalized confusion matrix for assessing area estimates from remotely sensed data. International journal of remote sensing, 22(16), 3223-3235.
Lin, T. Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D. & Zitnick, C. L. (2014). Microsoft coco: Common objects in context. In European conference on computer vision (pp. 740-755). Springer, Cham.
Liu, W., Yang, M., Xie, M., Guo, Z., Li, E., Zhang, L., ... & Wang, D. (2019). Accurate building extraction from fused DSM and UAV images using a chain fully convolutional neural network. Remote Sensing, 11(24), 2912.
Liu, Y., Zhang, Z., Liu, X., Wang, L. & Xia, X. (2021). Efficient image segmentation based on deep learning for mineral image classification. Advanced Powder Technology, 32(10), 3885-3903.
Long, J., Shelhamer, E. & Darrell, T. (2015). Fully convolutional networks for semantic segmentation. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition.
OGC. (2021). OGC city geography markup language (CityGML) 3.0 conceptual model users guide. In: Open Geospatial Consortium (OGC).
Ongsulee, P. (2017). Artificial intelligence, machine learning and deep learning. In 2017 15th international conference on ICT and knowledge engineering (ICT&KE) (pp. 1-6). IEEE.
Pan, Z., Xu, J., Guo, Y., Hu, Y. & Wang, G. (2020). Deep learning segmentation and classification for urban village using a worldview satellite image based on U-Net. Remote Sensing, 12(10), 1574.
Persello, C. & Stein, A. (2017). Deep fully convolutional networks for the detection of informal settlements in VHR images. IEEE Geoscience and Remote Sensing Letters, 14(12), 2325-2329.
Redmon, J., Divvala, S., Girshick, R. & Farhadi, A. (2016). You only look once: Unified, real-time object detection. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition.
Ren, H., Xu, C., Ma, Z. & Sun, Y. (2022). A novel 3D-geographic information system and deep learning integrated approach for high-accuracy building rooftop solar energy potential characterization of high-density cities. Applied Energy, 306, 117985.
Ronneberger, O., Fischer, P. & Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. Paper presented at the Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18.
Schnabel, R., Wessel, R., Wahl, R. & Klein, R. (2008). Shape recognition in 3d point-clouds.The International Journal Of Robotics Research.
Schonberger, J. L. & Frahm, J.-M. (2016). Structure-from-motion revisited. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition.
Tang, L., Li, L., Ying, S. & Lei, Y. (2018). A full level-of-detail specification for 3D building models combining indoor and outdoor scenes. ISPRS International Journal of Geo-Information, 7(11), 419.
Tarsha-Kurdi, F., Landes, T. & Grussenmeyer, P. (2008). Extended RANSAC algorithm for automatic detection of building roof planes from LiDAR data. The photogrammetric journal of Finland, 21(1), 97-109.
Verdie, Y., Lafarge, F. & Alliez, P. (2015). LOD generation for urban scenes. ACM Transactions on Graphics, 34(ARTICLE), 30.
Wang, C., Zhang, Y., Cui, M., Ren, P., Yang, Y., Xie, X., . . . Xu, W. (2022). Active boundary loss for semantic segmentation. Paper presented at the Proceedings of the AAAI Conference on Artificial Intelligence.
Wang, L., Li, R., Duan, C., Zhang, C., Meng, X. & Fang, S. (2022). A novel transformer based semantic segmentation scheme for fine-resolution remote sensing images. IEEE Geoscience and Remote Sensing Letters, 19, 1-5.
Wang, L., Li, R., Zhang, C., Fang, S., Duan, C., Meng, X. & Atkinson, P. M. (2022). UNetFormer: A UNet-like transformer for efficient semantic segmentation of remote sensing urban scene imagery. ISPRS Journal of Photogrammetry and Remote Sensing, 190, 196-214.
Westoby, M. J., Brasington, J., Glasser, N. F., Hambrey, M. J. & Reynolds, J. M. (2012). ‘Structure-from-Motion’photogrammetry: A low-cost, effective tool for geoscience applications. Geomorphology, 179, 300-314.
Xu, J., Zeng, F., Liu, W. & Takahashi, T. (2022). Damage Detection and Level Classification of Roof Damage after Typhoon Faxai Based on Aerial Photos and Deep Learning. Applied Sciences, 12(10), 4912.
Xu, J., Zeng, F., Liu, W. & Takahashi, T. (2022). Damage Detection and Level Classification of Roof Damage after Typhoon Faxai Based on Aerial Photos and Deep Learning. Applied Sciences, 12(10), 4912.
Xu, N., Luo, J., Zuo, J., Hu, X., Dong, J., Wu, T., . . . Liu, H. (2020). Accurate suitability evaluation of large-scale roof greening based on RS and GIS methods. Sustainability, 12(11), 4375.
Yang, L., Driscol, J., Sarigai, S., Wu, Q., Chen, H. & Lippitt, C. D. (2022). Google earth engine and artificial intelligence (ai): a comprehensive review. Remote Sensing, 14(14), 3253.
Ying, Y., Koeva, M., Kuffer, M., Asiama, K. O., Li, X. & Zevenbergen, J. (2020). Making the third dimension (3D) explicit in hedonic price modelling: A case study of Xi’an, China. Land, 10(1), 24.
Zhang, B. & Zhang, L. (2017). Study on the application of 3D modeling based on UAV photography in urban planning—Taking Yi Jiequ area in Du Jiangyan as an example. Paper presented at the AIP Conference Proceedings.
不動產估價師公會. (2018). 不動產估價師月刊107年12月. Retrieved from http://www.reaa.org.tw/downloads/file_management/10000/1000/66/20190130162733_66.pdf
中央氣象局. (2020). 109年氣候年報. Retrieved from https://www.cwb.gov.tw/Data/service/notice/download/Publish_20210610120207.pdf
王淑芬.(2019). 美濃水稻倒伏災損農委會首次啟動無人機勘災. 檢自https://www.cna.com.tw/news/firstnews/201905210277.asp
白肇亮. (1987). 台灣地區建築物屋頂構法之研究. (碩士). 國立成功大學建築(工程)研究所論文.
江映青. (2016). 用虛擬世界讓真實世界更美好. 國立成功大學機構典藏.
吳瑞賢, 蘇文瑞, 廖偉民, & 張志誠. (2002). 臺灣的颱風, 暴雨災害量化分析. 農業氣象及農業水資源之應用與管理.
汪知馨 & 邱式鴻. (2022). 以深度學習萃取高解析度無人機正射影像之農地重劃區現況資訊. Journal of Photogrammetry and Remote Sensing, 27(4), 193-211.
周巧盈, 巫思揚, & 陳琦玲. (2018). 應用無人飛機航拍影像協助農業勘災—以香蕉災損影像判釋為例. Journal of Photogrammetry and Remote Sensing, 23(2), 83-101.
周巧盈, 巫思揚, & 陳琦玲. (2020). 無人飛行載具之航拍影像應用於水稻倒伏災損判釋. 台灣農業研究, 69(1), 25-45.
林顯易 & 謝名豐. (2015). 工業 4.0 中的智慧機器人. 科儀新知(205), 12-20.
洪哲倫, 張志宏 & 林宛儒. (2019). 工業 4.0 與智慧製造的關鍵技術: 工業物聯網與人工智慧. 科儀新知(221), 19-25.
范慶龍. (2021). 監督式機器學習於土地覆蓋分類效益之研究. Journal of Taiwan Land Research, 24(1), 67-94.
國立中央大學太空及遙測研究中心. (2008). 發展影像高精度正射糾正相關技術及系統. 內政部國土測繪中心97年度結案報告.
國立中央大學空及遙測研究中心. (2019). 常用遙測衛星特性比較. Retrieved from https://www.csrsr.ncu.edu.tw/rsrs/SatelliteComparison.php
張智安 & 傅于洳. (2021). 應用深度學習於航照正射影像之房屋偵測. Journal of Photogrammetry and Remote Sensing, 26(4), 209-220.
張寶堂. (2019). 利用無人飛機系統航拍輔助土地複丈 . (碩士). 國立臺灣師範大學地理學系研究所論文.
梁鋆立. (2015). 應用遙測技術監測三蘆地區碳吸存量變化之研究. Retrieved from 新北市政府105年度自行研究報告.
陳立恒. (2018). 以影像邊緣線資訊優化真實正射影像. (碩士). 國立臺灣大學土木工程研究所論文.
陳良健 & 溫仁佑. (2006). 高解析力衛星影像真實正射改正及遮蔽區域補償. Journal of Photogrammetry and Remote Sensing, 11(3), 249-260.
楊明德, 許鈺群, 曾信鴻, & 曾偉誠. (2019). 無人機於精準農業之應用. 科儀新知(220), 20-39.
楊明德, 陳柏安, 陳怡璇, 張巧琳, & 賴明信. (2021). 多時期無人機影像於稻作植株高度量測之應用. Journal of Photogrammetry and Remote Sensing, 26(4), 193-207.
楊明德. (2015). 無人機鑑災公平又快速. Retrieved from https://www.gvm.com.tw/article/21160
趙丰. (2012). 太空垃圾知多少. Retrieved from https://www.ltedu.com.tw/web/scientific-epaper-content.aspx?KEY=143&ARTICLE=02.
電子全文 電子全文(網際網路公開日期:20280727)
連結至畢業學校之論文網頁點我開啟連結
註: 此連結為研究生畢業學校所提供,不一定有電子全文可供下載,若連結有誤,請點選上方之〝勘誤回報〞功能,我們會盡快修正,謝謝!
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
無相關期刊