災害防治為現今政府施政的方針之一。近年由於台灣西南沿海存在持續地層下陷的趨勢，導致排水不良的情況格外嚴重，故妥善的規劃建設該處之水利設施便為首要目標，以降低水患對環境所產生可能的衝擊。基於大範圍與高效率考量，本計畫擬整合遙測技術進行西南沿海區域監測，配合其他子計畫所產生之多時期數值地形模型，分析土地利用及地層下陷之關係。另，亦建置數碼城市與虛擬實境系統模擬災害的影響嚴重性，以提供決策者直覺式的參考依據。本計畫執行時間共計三年，第一年為針對目標區彙整蒐集各時期的遙測資料與相關文獻，進行土地利用與覆蓋類別判視，特別針對魚塭區域進行變遷偵測，觀察該區各時期的土地利用變化情況，分析與現有地層沈陷量之關聯性。第二年藉子計畫一蒐集與解算之航照立體對，與子計畫二之空載光達資料重建三維房屋模型，供建置三維數碼城市之用。第三年建立虛擬實境系統，展示三維數碼城市，並應用於災害境況模擬與地層下陷趨勢分析。研究重點包含水域偵測、房屋重建與展示視覺化。工作項目為偵測各時期遙測影像水域位置，並判別變遷區域；整合光達資料與航照立體對產生房屋輪廓線與高程，進而模塑三維房屋模型；開發虛擬實境系統模擬分析災害影響程度。 ; Hazard mitigation is one of the important government’s policies. Due to the subsidence that continuously lasts along the South-West coast of Taiwan, the drainage system is insufficient for flood prevention. Therefore, the main goal is to set up a well plan and perform quality construction for the proper drainage system to minimize hazard impacts. Based on the large coverage and high performance of remote sensing data, this project fuses the geoinformatic technologies to monitor the target areas. To match with the multi-temporal DEMs generated by other subprojects, we extract landuse information for the spatial analyses between the landuse and subsidence. In addition, we also establish Cyber City in a virtual reality environment for disaster simulation to provide better decision support. This project is planned to proceed for three years. In the first year, the collection of multi-temporal remote sensing data will be done for landuse change detection. The main target is on fish farms locating in the test areas. The major data sources are the satellite images over the selected time periods. Second, the orientation parameters of selected aerial images are calculated from sub-project one. Thus, the stereo photos and LIDAR data are used to reconstruct 3D building models for Cyber City. Additionally, the models are also mapped with texture from aerial images. Cyber City is then visualized for disaster simulation with virtual reality technology in the third year. The project cores include change detection and building reconstruction. The work items contain: (1) detection of water body over decades, (2) establish shapes and heights of 3D buildings from stereo photos and LIDAR data, (3) develop the virtual reality interface for disaster simulation. ; 研究期間 9708 ~ 9807
