臺灣本島擁有超過1200公里的海岸線，許多港口、遊憩區、水產養殖皆座落於此，且蘊含了豐富的生態系統。然而從地理資訊的角度而言，對於該區段的地形掌握長期以來仍屬欠缺，眾多全球數值高程模型（Digital Elevation Model, DEM）在海陸交界處皆呈現明顯不連續之高程，進而對於海岸變遷研究與船隻航行等造成困難。例如Shuttle Radar Topography Mission, SRTM）雖具有30公尺解析度並提供了全球陸地的高程資訊，但因雷達信號無法穿透水面，再加上拍攝時的無法與低潮的時間重疊，因此該產品對於沿岸的覆蓋範圍有限。此外，世界各地的沿海地區受到侵淤影響而變動快速，高成本的現地聲納量測缺乏即時性與全面性。因此，於海岸線快速更新數值高程模型，並用以監測人為及自然力量所導致的持續變化，已成為國土規劃和永續發展的關鍵任務。因此，本研究目標為延伸DEM的覆蓋範圍至低潮位陸側，並且劃設高低潮區間的潮間帶範圍。研究流程主要分為資料分析、重建沿岸地形、驗證結果以及潮間帶劃設。首先收集2010 – 2016這七年來的歷史無雲圖像，其中包括Landsat-5/-7/-8和SPOT-4/-5/-6/-7等超過10,000張影像，並計算每一張衛星影像的改良常態差異水體指標(Modified Normalized Difference Water Index , MNDWI)或常態差異水體指標(Normalized Difference Water Index , NDWI)來識別水體像素。再計算每一個影像網格中水所出現的機率並使用DTU10的潮汐模型做為高度參考，將淹水機率轉換為實際高程資訊。驗證部分則利用內政部地政司之海域基本圖的現地資料測試西部沿岸三區域（包含香山濕地、高美濕地以及外傘頂洲）地形。此外，於金門慈湖南側灘地另外運用無人飛行載具unmanned aerial vehicle (UAV)與建模軟體製作高程模型檢驗。本研究所製作之數值高程模型經UAV地形驗證可獲得約50公分精度，而由現地資料驗證後所得到root-mean-square of error (RMSE)介於37至62公分。除劃設全台潮間帶外，最後將以金門為例，運用不同時期所製作之海岸地形探討近岸沉積物的時空變化。

The coastline of Taiwan is over 1200 km and there are unique environment and landscapes such as ports, recreation areas, and aquacultures. However, this region is difficult to obtain the geo-information of the elevation because many Global Elevation Models (DEMs) have obviously discontinuous elevations at the land and sea and it caused it is difficult to research about coastline changing. For example, the Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM) with 30 m resolution provides enormously useful information about global terrestrial surfaces. However, the product has limited coverage over coastal area due to the impermeability of radar signal over water and the lack of low-tide coincidence. In addition, the coastline was fast changed by the erosion around the world but the sonar needs high-cost to measure and it is not immediacy and comprehensiveness method. Hence, to build an elevation model near the coastline and to monitor the rapid changes owing to human/ natural forces has become a critical task for planning and development. Therefore, we aim to extend to coverage of SRTM DEM for the determination of intertidal zone and to monitor temporal changes along the entire coastline of Taiwan (>1200km). We firstly collect historical cloud-free images since the 1980s, including Landsat series and SPOT series, and then calculate the Modified Normalized Difference Water Index (MNDWI) to identify water pixels. After computing water appearance probability of each pixel, it is converted into actual elevation by introducing the DTU10 tide model for high tide and low tide boundaries. For validation, we use the in-situ data from Ministry of the Interior to validate the three regions including Hsiang-Shan wetland, Gaomei wetland and Waisanding at west of Taiwan. In addition, we use the unmanned aerial vehicle (UAV) and Professional drone mapping and photogrammetry software (Pix4D) to validate the southern tidal flat of Ci Lake at Kinmen. In this research, the coastal DEM of intertidal zone is reconstructed and the accuracy is at 50 cm level as compared with in situ DEM built by an unmanned aerial vehicle (UAV). Moreover, the root-mean-square of error (RMSE) is between 37 to 62 cm as compared with in-situ DEM data from Ministry of the Interior. Finally, we give an example of Kinmen area to estimate temporal changes of nearshore sediment.

摘要 I
Abstract III
致謝 V
1. Introduction 1
1.1 Background and Motivation 1
1.2 Remote Sensing of Water 2
1.3 Objective 3
1.4 Architecture 4
2. Related Work 5
2.1 Applying Satellite Imageries to Monitor Coastline Changes 5
2.2 Using Multispectral Images to Detect Water Body. 7
2.3 Constructing the Coastal/Tidal Flat Digital Elevation Model (DEM) 8
2.4 Apply the Unmanned aerial vehicle (UAV) to construct 9
3. Study Area 10
3.1 Coastline of Taiwan 10
3.2 Coastline of Three Outlying Islands 11
4. Data and Methodology 13
4.1 Optical Remote Sensing Imageries 14
4.1.1 Landsat series 14
4.1.2 SPOT series 19
4.1.3 Sentinel-2 20
4.2 Image processing 22
4.3 DTU10 26
5. Results 29
5.1 Costal DEM of Taiwan 29
5.2 Delineating Intertidal zone 34
5.3 Validation 36
5.3.1 In situ data 36
5.3.2 Unmanned Aerial Vehicle (UAV) 39
5.4 Quantifying Coastline Changes in Kinmen 43
6. Discussions 46
7. Conclusions 50
8. Future work 51
8.1 Increase the resolution 51
8.2 Increase the accuracy of identification of pixels 52
References 54

[1] Stutz, M. L., & Pilkey, O. H. (2002). Global distribution and morphology of deltaic barrier island systems. Journal of Coastal Research, 36(sp1), 694-707.
[2] Athearn, N. D., Takekawa, J. Y., Jaffe, B., Hattenbach, B. J., & Foxgrover, A. C. (2010). Mapping elevations of tidal wetland restoration sites in San Francisco Bay: comparing accuracy of aerial lidar with a singlebeam echosounder. Journal of Coastal Research, 312-319.
[3] Teodoro, A. C. (2016). Optical Satellite Remote Sensing of the Coastal Zone Environment—An Overview. In Environmental Applications of Remote Sensing. InTech.
[4] Siripong, A. (2010). Detect the coastline changes in Thailand by remote sensing. International Archives of the Photogrammetry, Remote Sensing and Spatial Information Science, 38(Part 8), 992-996.
[5] Ekercin, S. (2007). Coastline change assessment at the Aegean Sea coasts in Turkey using multitemporal Landsat imagery. Journal of Coastal Research, 691-698.
[6] Bouchahma, M., Yan, W., & Ouessar, M. (2012). Island coastline change detection based on image processing and remote sensing. Computer and Information Science, 5(3), 27.
[7] Li, R., Liu, J. K., & Felus, Y. (2001). Spatial modeling and analysis for shoreline change detection and coastal erosion monitoring. Marine Geodesy, 24(1), 1-12.
[8] Wang, Z., Gao, S., & Huang, H. (2010). Spatial variations of tidal water level and their impact on the exposure patterns of tidal land on the central Jiangsu coast. Acta Oceanologica Sinica, 29(1), 79-87.
[9] McFeeters, S. K. (1996). The use of the Normalized Difference Water Index (NDWI) in the delineation of open water features. International journal of remote sensing, 17(7), 1425-1432.
[10] Xu, H. (2006). Modification of normalised difference water index (NDWI) to enhance open water features in remotely sensed imagery. International journal of remote sensing, 27(14), 3025-3033.
[11] Mason, D. C., Davenport, I. J., Robinson, G. J., Flather, R. A., & McCartney, B. S. (1995). Construction of an inter‐tidal digital elevation model by the ‘Water‐Line’Method. Geophysical Research Letters, 22(23), 3187-3190.
[12] Webster, T. L., & Forbes, D. L. (2005, November). Using airborne LIDAR to map exposure of coastal areas in Maritime Canada to flooding from storm-surge events: A review of recent experience. In Proceedings of Canadian Coastal Conference, Dartmouth, Canada (Vol. 69).
[13] Tseng, K. H., Kuo, C. Y., Lin, T. H., Huang, Z. C., Lin, Y. C., Liao, W. H., & Chen, C. F. (2017). Reconstruction of time-varying tidal flat topography using optical remote sensing imageries. ISPRS Journal of Photogrammetry and Remote Sensing, 131, 92-103.
[14] Yen, C. H., Chen, K. T., Lee, S. P., Liu, C. J., Wu, C. Y., & Chan, H. C. (2015). A feasibility study on unmanned aerial vehicle for river stability. Journal of soil and water conservation, 47(3), 1407-1416. (Chinese)
[15] Pai, C. C., Liu, Y. C., Hsiao, Y. S., Lien, H. P., & Lin, P. H. (2015). Analysis of Accuracy for UAV-derived Topography from a GoPro Camera. Journal of Chinese soil and water conservation, 46(3), 142-149. (Chinese)
[16] Stephenson, N. M. (2016). High resolution habitat suitability modeling for a narrow-range endemic alpine Hawaiian species (Doctoral dissertation, University of Hawai'i at Hilo).
[17] Terrain of Taiwan coastal: http://www.tlsh.tp.edu.tw/~t127/topographytaiwan/
[18] Ministry of the Interior, Republic of China (Taiwan): https://www.cpami.gov.tw/filesys/file/chinese/dept/rp3/rp1060206.pdf (Chinese)
[19] Penghu information: https://www.penghu-nsa.gov.tw/InDepthExperience/AboutUs/AboutPenghu01.htm
[20] Earthexplorer website: https://earthexplorer.usgs.gov/
[21] Wulder, M. A., White, J. C., Loveland, T. R., Woodcock, C. E., Belward, A. S., Cohen, W. B., ... & Roy, D. P. (2016). The global Landsat archive: Status, consolidation, and direction. Remote Sensing of Environment, 185, 271-283.
[22] The Landsat-8 product: https://landsat.usgs.gov/using-usgs-landsat-8-product
[23] Henry, P., & Meygret, A. (2001). Calibration of HRVIR and VEGETATION cameras on SPOT4. Advances in Space Research, 28(1), 49-58.
[24] ESA (European Space Agency). 2015a. “Sentinel-2 User Handbook, Issue 1, Rev 2, Revision 2.” ESA Standard Document. Accessed July 24 2015, 64 pages.
[25] Cheng, Y., & Andersen, O. B. (2010). Improvement in global ocean tide model in shallow water regions. Poster, SV, 1-68.
[26] Cheng, Y., & Andersen, O. B. (2011). Multimission empirical ocean tide modeling for shallow waters and polar seas. Journal of Geophysical Research: Oceans, 116(C11).
[27] DTU website: http://www.space.dtu.dk/english/
[28] Hsu, Robert & Chang, K.C. (2015). The Use of Innovative Software Pix4DMapper to Optimize the Process of Generating Spatial Data from UAV’S Aerial Images. Journal of the Chinese Institute of Civil and Hydraulic Engineering, Vol. 27, No. 3, pp. 241–246. (Chinese)
[29] DJI website: https://store.dji.com/
[30] ArcGis Pro Tool Reference: http://pro.arcgis.com/en/pro-app/tool-reference
[31] Landsat-8 QA band information: https://landsat.usgs.gov/collectionqualityband