傳統遙測技術受限於單一感測器的物理特性，在面對複雜都市環境時，常因垂直視角受限、光譜混淆與資料時效落差，導致顯著的觀測盲點。為突破此一技術瓶頸，本研究提出一套整合時間、空間與物理量的多源遙測辨識框架，並以具備高隱蔽性與材質混淆特性的石綿瓦作為實證案例，驗證此方法論在環境治理上的有效性。

本研究方法分為三層次整合：在物理量整合上，結合航拍影像的紋理特徵（CNN）與衛星光譜特徵（SVM），解決材質混淆問題；在空間整合上，建構「由上而下（衛星/航拍）」與「由下而上（街景/公民參與）」的雙模組架構，利用YOLOv9-E模型克服立面遮蔽死角；在時間整合上，引入Sentinel-1雷達影像進行時序檢核，以排除已拆除之建物。

研究結果顯示，本框架成功完成了全台灣共計292,876棟含石綿瓦屋頂建築物的定位與測繪。空間整合分析證實，有60.4%的含石綿建物僅能透過地面視角檢出，量化證明了多視角整合的必要性。此外，經由現地採樣修正密度參數（2,356 kg/m³）後，推估全台潛在廢棄物總量約54.5萬公噸。最後，本研究證實結合AI輔助的公民參與模式（F1-Score 90.91%）能有效補足自動化街景的偵測盲區。

本研究成功建立了一套從宏觀普查到一般尺度辨識的全方位偵測流程。研究成果不僅產出了全國首份建物級石綿分佈圖資，更實證了多源遙測整合能有效突破單一視角的物理限制，可作為未來環境監測與國土治理的技術典範。

Traditional remote sensing is often constrained by the physical limitations of single sensors, leading to significant observational blind spots in complex urban environments due to restricted vertical viewing angles, spectral confusion, and temporal data gaps. To overcome these technical bottlenecks, this study proposes an integrated identification framework incorporating temporal, spatial, and physical dimensions. Asbestos tiles, characterized by high concealment and material ambiguity, were selected as an empirical case to validate the effectiveness of this methodology for environmental governance.

The methodology involves three levels of integration: (1) Physical Integration: Combining aerial texture features (CNN) with satellite spectral signatures (SVM) to resolve material confusion; (2) Spatial Integration: Establishing a dual-module architecture that merges "top-down" (satellite/aerial) and "bottom-up" (street view/citizen science) perspectives, utilizing the YOLOv9-E model to overcome facade occlusion; and (3) Temporal Integration: Introducing Sentinel-1 radar imagery for sequential verification to exclude demolished buildings.

The results indicate that the framework successfully mapped 292,876 asbestos-roofed buildings across Taiwan. Spatial integration analysis revealed that 60.4% of asbestos buildings were detectable only through ground-level perspectives, quantitatively confirming the necessity of multi-view integration. Furthermore, by correcting density parameters via field sampling (2,356 kg/m³), the total potential hazardous waste was estimated at 545,570 tons. Finally, the study demonstrated that the AI-assisted citizen science approach (F1-Score 90.91%) effectively filled the detection gaps left by automated street view analysis.

This study successfully establishes a comprehensive detection process ranging from macroscopic surveys to microscopic identification. The findings not only produced the first nationwide building-level asbestos database but also empirically verified that multi-source remote sensing integration can effectively surmount the physical limitations of single perspectives, serving as a technical paradigm for future environmental monitoring and land governance.

內政部 (2011)。建築技術規則建築構造編。內政部營建署。
Abbasi, M., Hosseiny, B., Stewart, R. A., Kalantari, M., Patorniti, N., Mostafa, S., & Awrangjeb, M. (2024). Multi-temporal change detection of asbestos roofing: A hybrid object-based deep learning framework with post-classification structure. Remote Sensing Applications: Society and Environment, 34, 101167. https://doi.org/10.1016/j.rsase.2024.101167
Aranda, N. R., De Waegemaeker, J., & Van de Weghe, N. (2023). The evolution of public participation GIS (PPGIS) barriers in spatial planning practice. Applied Geography, 155, 102940. https://doi.org/10.1016/j.apgeog.2023.102940
Baek, S.-C., Lee, K.-H., Kim, I.-H., Seo, D.-M., & Park, K. (2023). Construction of Asbestos Slate Deep-Learning Training-Data Model Based on Drone Images. Sensors, 23(19), 8021. https://doi.org/10.3390/s23198021
Bassani, C., Cavalli, R. M., Cavalcante, F., Cuomo, V., Palombo, A., Pascucci, S., & Pignatti, S. (2007). Deterioration status of asbestos-cement roofing sheets assessed by analyzing hyperspectral data. Remote Sensing of Environment, 109(3), 361-378. https://doi.org/10.1016/j.rse.2007.01.014
Campopiano, A., Ramires, D., Zakrzewska, A. M., Ferri, R., D'annibale, A., & Pizzutelli, G. (2009). Risk Assessment of the Decay of Asbestos Cement Roofs. The Annals of Occupational Hygiene, 53(6), 627-638. https://doi.org/10.1093/annhyg/mep036
Cilia, C., Panigada, C., Rossini, M., Candiani, G., Pepe, M., & Colombo, R. (2015). Mapping of asbestos cement roofs and their weathering status using hyperspectral aerial images. ISPRS International Journal of Geo-Information, 4(2), 928-941. https://doi.org/10.3390/ijgi4020928
Doll, R. (1955). Mortality from lung cancer in asbestos workers. British journal of industrial medicine, 12(2), 81. https://doi.org/10.1136/oem.12.2.81
Fatima, Z., Tanveer, M. H., Mariam, H., Voicu, R. C., Rehman, T., & Riaz, R. (2024). Performance Comparison of Object Detection Models for Road Sign Detection Under Different Conditions. International Journal of Advanced Computer Science & Applications, 15(12), 996-1005. https://doi.org/10.14569/ijacsa.2024.0151299
Frassy, F., Candiani, G., Rusmini, M., Maianti, P., Marchesi, A., Nodari, F. R., Via, G. D., Albonico, C., & Gianinetto, M. (2014). Mapping asbestos-cement roofing with hyperspectral remote sensing over a large mountain region of the Italian western alps. Sensors, 14(9), 15900-15913. https://doi.org/10.3390/s140915900
Gibril, M. B. A., Shafri, H. Z., & Hamedianfar, A. (2017). New semi-automated mapping of asbestos cement roofs using rule-based object-based image analysis and Taguchi optimization technique from WorldView-2 images. International Journal of Remote Sensing, 38(2), 467-491. https://doi.org/10.1080/01431161.2016.1266109
Grizonnet, M., Michel, J., Poughon, V., Inglada, J., Savinaud, M., & Cresson, R. (2017). Orfeo ToolBox: Open source processing of remote sensing images. Open Geospatial Data, Software and Standards, 2, 1-8. https://doi.org/10.1186/s40965-017-0031-6
Gubiani, D., Sgrazzutti, G., Basso, M., Viero, E., Tavaris, D., Foresti, G. L., & Scagnetto, I. (2025). A dynamic neural network model for the identification of asbestos roofings in hyperspectral images covering a large regional area. Computer‐Aided Civil and Infrastructure Engineering, 40(7), 879-894. https://doi.org/10.1111/mice.13376
Hasan, H., Shafri, H. Z., & Habshi, M. (2019). A comparison between support vector machine (SVM) and convolutional neural network (CNN) models for hyperspectral image classification. IOP Conference Series: Earth and Environmental Science, 357(1), 012035. https://doi.org/10.1088/1755-1315/357/1/012035
Kiviluoto, R. (1960). Pleural calcification as a reontgenologic sign of non-occupational endemic anthophyllite-asbesto-sis. Acta radiol, 194, 1-67.
Krówczyńska, M., Raczko, E., Staniszewska, N., & Wilk, E. (2020). Asbestos—cement roofing identification using remote sensing and convolutional neural networks (CNNs). Remote Sensing, 12(3), 408. https://doi.org/10.3390/rs12030408
Krówczyńska, M., & Wilk, E. (2019). Environmental and occupational exposure to asbestos as a result of consumption and use in Poland. International Journal of Environmental Research and Public Health, 16(14), 2611. https://doi.org/10.3390/ijerph16142611
Krówczyńska, M., Wilk, E., Pabjanek, P., & Kycko, M. (2017). Hyperspectral Discrimination of Asbestos‑Cement Roofing. Geomatics and Environmental Engineering, 11(1), 47-65. https://doi.org/10.7494/geom.2017.11.1.47
Lanaras, C., Bioucas-Dias, J., Galliani, S., Baltsavias, E., & Schindler, K. (2018). Super-resolution of Sentinel-2 images: Learning a globally applicable deep neural network. ISPRS Journal of Photogrammetry and Remote Sensing, 146, 305-319. https://doi.org/10.1016/j.isprsjprs.2018.09.018
Lin, Y.-C., Yu, T.-T., Yang, Y.-E., Lin, J.-C., Lien, G.-W., & Lan, S.-C. (2025). Merging Various Types of Remote Sensing Data and Social Participation GIS with AI to Map the Objects Affected by Light Occlusion. Remote Sensing, 17(13), 2131. https://doi.org/10.3390/rs17132131
Lynch, K. M., & Smith, W. A. (1935). Pulmonary asbestosis III: Carcinoma of lung in asbesto-silicosis. The American Journal of Cancer, 24(1), 56-64.
Pacifici, F. (2016). On the predictive value of the WorldView3 VNIR and SWIR spectral bands. 2016 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), 898-901. https://doi.org/10.1109/IGARSS.2016.7729227
Pinho, C. M. D. d., Silva, F., Fonseca, L., & Monteiro, A. (2008). Intra-urban land cover classification from high-resolution images using the C4. 5 algorithm. ISPRS Congress Beijing, 7. https://doi.org/10.1080/01431161.2012.675451
Spurny, K. (1989). On the release of asbestos fibers from weathered and corroded asbestos cement products. Environmental research, 48(1), 100-116. https://doi.org/10.1016/s0013-9351(89)80089-1
Szabó, S., Burai, P., Kovács, Z., Szabó, G., Kerényi, A., Fazekas, I., Paládi, M., Buday, T., & Szabó, G. (2014). Testing algorithms for the identification of asbestos roofing based on hyperspectral data. Environmental Engineering and Management Journal, 143(11), 2875-2880. https://doi.org/10.30638/eemj.2014.323
Tishby, N., & Zaslavsky, N. (2015). Deep learning and the information bottleneck principle. 2015 ieee information theory workshop (itw), 1-5. https://doi.org/10.1109/ITW.2015.7133169
Tommasini, M., Bacciottini, A., & Gherardelli, M. (2019). A QGIS tool for automatically identifying asbestos roofing. ISPRS International Journal of Geo-Information, 8(3), 131. https://doi.org/10.3390/ijgi8030131
Wang, C.-Y., Yeh, I.-H., & Mark Liao, H.-Y. (2024). Yolov9: Learning what you want to learn using programmable gradient information. European conference on computer vision, 1-21. https://doi.org/10.1007/978-3-031-72751-1_1
Wu, T., Wu, X.-J., & Luo, X.-Q. (2012). A study on fusion of different resolution images. Procedia engineering, 29, 3980-3985. https://doi.org/10.1016/j.proeng.2012.01.605
Yu, T.-T., Lin, Y.-C., Lan, S.-C., Yang, Y.-E., Wu, P.-Y., & Lin, J.-C. (2022). Mapping asbestos-cement corrugated roofing tiles with imagery cube via machine learning in Taiwan. Remote Sensing, 14(14), 3418. https://doi.org/10.3390/rs14143418
Zou, Z., Chen, K., Shi, Z., Guo, Y., & Ye, J. (2023). Object detection in 20 years: A survey. Proceedings of the IEEE, 111(3), 257-276. https://doi.org/10.1109/Jproc.2023.3238524