衛星影像於台灣地區淹水範圍判釋應用之可行性評估

The Feasibility of Flood Boundary Delineation Using Remote Sensing Images in Taiwan

江介倫(Jie-Lun Chiang) ； 劉俊志(Jun-Jih Liou)

《科學與工程技術期刊》 7卷4期 (2011/12) Pp. 1-8

https://doi.org/10.7117/JSET.201112.0003

摘要

台灣地區，水的問題日益嚴重，尤其在颱風季節，洪澇災害等淹水問題更是常見，國外已有很多研究成功的應用衛星遙測技術即時或近即時監測淹水範圍。惟台灣河川大多坡陡流急，以致洪水之停滯時間短，淹水常於一日內，甚或數小時間即消退，故要於淹水時段取得衛星影像不易，再者台灣地狹人稠土地利用複雜，亦造成由衛星影像判釋淹水區域之困難，唯衛星影像可監控大範圍地表狀況，且主動式影像受天候影像較小的優點，在此類應用仍頗具發展潛力。因此本研究透過衛星影像廣景攬要的優點，蒐集納莉、安珀等歷史颱洪事件前後之衛星影像資料，以獲取當時地面資訊，並加以分析。藉由影像分類與不同時期影像之變遷偵測等方式分析由衛星影像紀錄到的歷史淹水事件之淹水區域與範圍，期能提供淹水時之地表資訊，以提供如防救災或保險理賠等相關單位參考，甚或提供淹水模式之驗證資料等後續應用。

關鍵字

衛星遙測 ； 淹水 ； 颱洪 ； 分類

並列摘要

In Taiwan, water is a serious concern. During typhoon season, water-related natural disasters, such as landslides and floods, also occur. Although other countries have successfully employed remote sensing to monitor flooding in real-time, flood monitoring is substantially more difficult in Taiwan. This is primarily because of the country's steep landforms and the limited flood water stagnation period; flood waters typically recede within a day or even several hours. Therefore, obtaining satellite images during the flood period is difficult. Furthermore, Taiwan's limited land area, high population density, and multiple land uses hinder differentiation between the flood areas in satellite images. However, satellite imagery has the advantage of enabling large-scale monitoring of land surfaces and is less influenced by weather conditions (active images), which is essential for this field. Records of flood areas are not easy to obtain. Additionally, physically investigating large-scale areas during flooding or after flooding is dangerous and costly. This study obtains and analyzes satellite images of previous typhoons, such as Nari and Amber, and of flooded areas taken during previous floods. The results of this study can enhance flood information and provide a useful reference for rescue and insurance organizations. The results can also provide data for verifying flood models and related studies.

並列關鍵字

satellite remote sensing ； flood ； typhoon ； classification