大型邊坡滑動經常造成嚴重的生命及財產損失，其中順向坡滑動破壞相較其他類型邊坡滑動所造成的災害更為巨大，因此，潛在順向坡圈繪成為防範順向坡滑動的重要前置工作之一。順向坡的圈繪工作，早期是由人工判釋繪製而成，圈繪結果往往較為主觀且費時。現今因遙測技術的進步以及電腦運算效能的提升，使得順向坡圈繪可能可以透過電腦程式自動化判釋，使得圈繪結果能較為客觀且效率提升。 本研究採用高解析度的數值高程模型(DEM)及既有的地質位態兩類資料作為已知條件輸入至地理資訊系統(GIS)，給定符合順向坡之地質條件，並透過一系列自動化流程進行順向坡圈繪。本研究首先將既有地層位態資料，分別透過趨勢面法及克里金空間內插法轉換得到GIS地層傾向位態圖層，再將兩方法所得之傾向圖層與DEM圖層資料一同進行網格疊加分析，得出符合順向坡條件之網格圖層，再以點密度(Point Density)演算法自動圈繪順向坡，最後，本研究將自動化圈繪的順向坡結果與政府公告順向坡圖資以及已發生順向坡滑動範圍做比對，研究結果顯示，以克里金內插法所得位態圖層並圈繪之順向坡於本研究區域有較高的一致性。 此外，本研究以圈繪結果為基礎，進一步找出坡角大於傾角的網格分布，定義為「見光(Daylight)的危險順向坡」，並透過野外調查進行現地查核，結果顯示，現地多數危險順向坡大多已有地錨或擋土牆等防護設施，少部分邊坡有植生覆蓋不易調查，但整體成效良好，顯示本研究所提之高精度數值地形資料搭配自動化判釋步驟，可有效快速的找出潛在危險順向坡可能的分布。;Large landslides usually caused severe damage to people and properties. The disaster scales of dip-slope landslides were more severe than other types of landslides. Therefore, the dip-slope mapping becomes one of the important works for preventing landslides. According to Soil and Water Conservation Technical Regulations Article No. 31, the dip slopes are defined such that the angle between the strikes of the slope surface and the bedding or the angle between the strikes of the slope surface and the weak plane is less than 20 degrees with the same dip directions. In the past, dip-slope mapping methods were usually acquired by manual mapping. The results were subjective and time-consuming. Now, with the advanced development of remote-sensing technology and computing power, the dip-slope mapping could be automatically interpreted by fast computer program analysis. This technology makes dip-slope mapping more objective and efficient. The study proposes to input the high-resolution digital elevation model (DEM) and the attitudes of the bedding planes from field investigation into Geographic Information System (GIS). We intended to identify the ranges for dip slopes by executing series of automated identification program. First, the study uses GIS to obtain dip direction of bedding information raster data from the trend surface method and kriging by the actual bedding attitude data, respectively. Afterward, we overlaid two kinds of dip direction raster data respectively to obtain the dip-slope cell raster with aspect and slope raster data from DEM data. Then, we execute “Point-Density” algorithm to automatically finish the dip-slope mapping by choosing the appropriate Point Density (PD) value. Last, we take the results of dip-slope mapping to compare with the dip-slope raster data by CGS and the landslide cases in the study area. We consider that the attitude raster data making by kriging have higher accuracy than the other one in the study area. Based on the dip-slope mapping result in this study, we further consider the distribution of the cells of which slope angles are bigger than dip angles, and we called the involved dip-slopes the cells “the daylighted dip-slopes with potential hazard”. Finally, we try to explore the evidence about whether the mapped dip-slopes are danger dip-slopes by field investigation. In the future, we hope the automated program in our study can effectively and rapidly find the possible distribution of the dip-slopes with potential hazards, and also reduce the cost and time about the geological field investigation for dip slope mapping.
