海岸帶是海洋與大陸相互作用的地帶，具有高度自然能量和生物生產力。桃園海岸段長度約為46公里，海岸組成物質以沙、礫石及珊瑚礁為主。其海岸帶原本具有豐富多樣性的海岸生態，但在長時間人為及自然因素下，造成海岸線侵淤與人工構造物的變遷，因此對於保育區的監測以及地形變遷的了解是一項重要議題。本計劃將擬定一套以衛星資料(光學影像、雷達影像)為主的系統計算海岸地區實際高程，利用計算後得到的平均潮位線監測海岸線變化。透過收集歷年光學影像(例如Landsat系列)，計算每一張衛星影像的改良常態差異水體指標 (MNDWI)，或利用Sentinel-1雷達影像，透過後向散射的強度來辨識水體面積，隨後將所有影像計算出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。此外可再利用光學影像的光譜演算法計算出近岸水深，最終將兩種高程合併成桃園沿岸地形，可探討桃園海岸線近三十年的變化趨勢 ;Coastal area is the home to many conservation animals and its environmental protection is an important issue all over the world. Taoyuan coastline is about 46 km and the coastal constituents are mainly sand, gravel and coral reefs. However, natural and human factor have altered coastal terrain and threatened the ecological system in this area. Therefore, we develop a workflow that utilizes satellite imageries to track the long-term variation of the coastline. First, we collect multiple optical and radar satellite images, including Landsat series, Sentinel-1, and Sentinel-2, to identify water pixels. For optical satellite imageries, we calculate the Modified Normalized Difference Water Index (MNDWI) and for radar imageries we use the single-threshold of backscatter intensity for water classification. Next, we sum up all imageries to calculate water appearance probability of each pixel and convert it into actual elevation by introducing the DTU10 tide model. Moreover, we can use the optical satellite imageries to calculate the shallow water bathymetry and combine these two elevations to construct the Taoyuan coastal DEM and detect the trend of coastline changing.
