The creation of landslide inventories is an important step in landslide susceptibility mapping, and automated algorithms for landslide detection will increasingly be relied upon as part of the mapping process. This study compares the effects of three different DTM generation methods on a pixel-based landslide detection algorithm developed by Shi et al. (2018) using a set of landslide-prone study areas in Pierce County, Washington. Non-parametric statistical analysis demonstrated that false-positive and false-negative rates were significantly different between DTM generation methods, showing that inpainting presents a more balanced error profile compared to TIN and morphological-based approaches. However, overall accuracy (kappa) rates were still very low overall, suggesting that geomorphometric curvature as an input needs to be processed in a different manner to make these types of pixel-based landslide detection algorithms more useful for landslide inventory database management. / Master of Science / The creation of landslide inventories is an important step in landslide susceptibility mapping, and automated algorithms for landslide detection will increasingly be relied upon as part of the mapping process. This study compares the effects of three different DTM generation methods on a pixel-based landslide detection algorithm developed by Shi et al. (2018) using a set of landslide-prone study areas in Pierce County, Washington. Statistical analysis demonstrated that false-positive and false-negative rates were significantly different between DTM generation methods, showing that inpainting presents a more balanced error profile compared to TIN and morphological-based approaches. However, overall accuracy rates were still very low overall, suggesting that curvature as an input needs to be processed in a different manner to make these types of pixel-based landslide detection algorithms more useful for landslide inventory database management.

完善的崩塌地資料庫有助於地區土地利用的適宜性評估、與環境保護措施之研訂。目前，崩塌地地理空間資料(Geospatial data)的產生方法主要為：人為判釋經正射糾正(Ortho-rectification)的遙測影像，基於該影像，將辨識目標數位化(Digitizing)。然而，遙測影像的「正射糾正」與「人為判釋」往往不敷災後的緊急需求。為促進資料收集效率，本研究試圖發展一套自動化流程：以「未經糾正的遙測影像」為判釋對象，判釋作業以「物件式影像分類(Object-based classification)技術」進行，並利用「現存地形資料」，實現自動判釋所產生之辨識成果的地理對位(Georeferencing)與過濾篩選；最後，以「與現存各類輔助資料的套疊分析成果」為其屬性，以便利崩塌地地理空間資料的後續應用。 物件式影像分類分為為「影像分割(Image segmentation)」與「物件分類」兩步驟。於影像分割階段，採用多重解析度分割法(Multiresolution segmentation algorithm)─由於陰影下各類地物的影像光譜差異較不明顯，為避免陰影區之錯誤分割，賦予陰影區較小的尺度參數(Scale parameter)；於物件分類階段，基於訓練資料，以「線性核函數的支持向量機(Support Vector Machine, SVM, with a linear kernel)」為分類器，偵測「非雲與植被區」，並輸出為向量式資料(Vector data)。而後基於現存地形資料，以光線追蹤法(Ray-tracing algorithm)進行分類器輸出向量式資料的地理對位，並自訂第二階段的地形特徵過濾準則。實驗成果顯示，此自動化流程產出的崩塌地地理空間資料─其生產者精度(Producer’s accuracy)與使用者精度(User’s accuracy)分別介於0.85~0.99與0.44~0.96。