在空間資訊中所擷取屬性因子眾多諸如地形、植生和地貌，由於大量的資料被重複的使用，且資料中不確定性很高，以致在建構坡地災害實易造成誤判。再者，土石崩塌後的地貌演變，無法獲
得有效的預測分析。本研究首先使用不同時期SPOT 衛星影像所測得的資料來作為研究屬性，並抽樣資料建立不同時期的雪霸崩塌地的資料庫作為準確率分析依據，之後利用粗糙集離散化(DiscreteRough Set)理論，從眾多的屬性中，找出影響坡地變化的關鍵屬性以及分離門檻值，進而建構出使用細胞自動機技術，預測崩塌下次發生的位置，建構其崩塌地演化規則與推估迴歸公式，並考慮週圍網格對目標網格的影響，以提高評估的準確率。

Abstract Spatial information of landslide interpretation is usually analyzed by statistical analysis based on topography, vegetation and
landforms. However, due to the uncertainty of the data and the mass of
data, the determination will result in many misjudgments. The study starts
with adopting various scenario SPOT image data to develop the landslide
database. Then, an innovative Data Mining technique Discrete Rough Sets
(DRS) is applied to attain the core factors and the thresholds of the
multi-categories on image classification. In the mean time, integrating the thresholds of slopes, the classification of landslide zones of landslide,bare-land, stream, and water-body is greatly improved. In the second stage, the Cellular Automata with neighborhood characteristics is applied to obtain the influences of the terrain features. Various categories by logistic regression functions are used to calculate the probability of occurrence for each category. The DRS is also used as a classifier to attain the variation of landforms. The study found that using DRS can refine the outcomes of predictions but Cellular Automata can enhance the influence by the vicinity of grid-cell.