線形特徵是地表形狀中最為常見的形狀特徵，其中包括了山脈、河道、斷層、以及褶皺，影響這些線形的因素有很多，如地層的受力方向會影響山脈走向，地質的性質與受力的大小會形成斷層及褶皺，進而影響水系的分佈，因此藉由線形的分析，可以進一步分析統計山脈走向，提供在地質構造及水系分析的資訊。目前在線形萃取的研究中，主要還是依賴人工手繪的方式，經過繪圖員的判斷描繪線段，再對線段的數量與方向做統計及分析。近年來由於影像解析度的增加，資料量大增，以人工方式將會耗去大量的時間，再加上由於個人主觀因素，不同的繪圖員會因其經驗而描繪出不同的結果，缺乏客觀性的依據。本研究主要目的在於以電腦自動化方式，利用數位高層模型及合成孔徑雷達影像，快速客觀的萃取線形，取得較客觀的萃取結果，以提供地質學的學者參考。 ; Lineament features are the most common features on the surface of ground. Those features represent mountain ridges, rivers, faults and folds. The direction of forces onto the stratum causes the different direction of the mountain ridges. The geological properties and the strength of the forces will form the faults or folds. Therefore, the extracted lineaments can provide information for the geological structure. For the researches with lineament information nowadays, the lineament extraction is still based on visual inspection for the analysis of the amounts and directions of lineaments. Because of the recent improvements in spatial resolution and the increment of data amounts, artificial extraction becomes lack of efficiency. Additionally, the performances are also subjected to the experiences of the plotters. Therefore, in the project, we proposed a computer-aided automatic lineament extraction method for Digital Elevation Model (DEM) and Synthetic Aperture Radar (SAR) images to objectively extract lineaments, and provide useful information for researchers in geological research areas. ; 研究期間 9708 ~ 9807
