頁籤選單縮合
題名 | 玉山國家公園內陳有蘭溪上游集水區地貌變遷動態資料庫建立之研究=Establishment of Dynamic Database for Environmental Changes in Chen-Yu-Lan Creek Watershed, Yu-San National Park |
---|---|
作者 | 林俐玲; 陳禮仁; 陳文福; Lin, Li-ling; Chen, L. R.; Chen, Wen-fu; |
期刊 | 水土保持學報 |
出版日期 | 19980900 |
卷期 | 30:3 1998.09[民87.09] |
頁次 | 頁207-227 |
分類號 | 432.734 |
語文 | chi |
關鍵詞 | 地理資訊系統; 影像處理; 動態資料庫; 不規則三角網; Geographical information system; Image processing; Dynamic database; Triangle irregular network; |
中文摘要 | 玉山國家公園面積廣袤,除了原有之林道、步道及少數的交通幹道外,園區內道 路很少,交通不便,在環境管理上是一大障礙。本研究旨在應用遙測技術結合地理資訊系 統之應用,以嘗試尋求建立本研究區早期地貌變遷動態資料庫之方法,冀能對其集水區內 環境之管理提供參考。以GIS軟體PC ArcC/Info建立本區基本自然資源與環境資料庫之圖檔 。另以SPOT衛星影像,利用影像處理乾體PC ERDAS對1986年11月和1990年3月兩期之衛星影 像進行地表覆蓋分類並加以套疊,以找出地貌發生變遷地區之位址與大小,此對未來後續 之查詢與管理上可提供有效性及即時性之參考;至於其後期之變化則擬以另文探討之。茲 將研究結果列述如後: 一、由地理資訊系統建立之基本資料庫有水系、河川階序、等高線、地質等圖及由等高線 資料作成之不規則三角網並作成坡度圖及坡向圖,而衛星影像分類後之結果則成土地 利用圖。最後將網格型資料轉成向量型資料。 二、影像處理與分類之心得: 1.在作輻射校正時,選擇波長較短之紅光及綠光作為起始波段,則兩波段間之分類結果並 無顯著差異。 2.在分類時因研究區位於上游,河道為V型谷,沖蝕營力強,河寬窄,而SPOT影像之空間解 像力僅20公尺*20公尺,較粗,河流及裸露地均無法再予區分。 3.影像經處理與分類所得之圖層經套疊,即可圈出變遷之地區。所得主要之結果有:A.金 門峒斷崖有向上游侵蝕的情形,由影像分析可看出其間之差異,若比較不同年份間之期 距夠長的話,情形將更為清楚。B.由4年間之土地利用╱覆蓋圖型態之變遷結果顯示:林 地與草生地各減少1.29與2.76%,崩塌地則增加4.05%。 |
英文摘要 | Yu-San National Park is very large in area. Besides forest road, footpaths and few major roads, access is not convenient within the park area. These reasons make environmental management very difficult. The purpose of the study is to build dynamic database with remote sensing techniques and application of GIS on cover changes for proper environmental management. GIS software PC ARC/INFO is used to establish the graphical files of database on natural resource and environment. On the other hand, the image process software PC ERDAS was used to classify the land cover from French satellite SPOT of 1986 and 1990 to find out the sites and sizes of land use changes. The results are summarized as follows: 1. GIS in database establishment: The basic data established in this study is river systems, river order, contour line, geological maps and the slope map and aspect map obtained from contour line and by using TIN(Triangle Irregular Network) module of ARC/INFO. The land cover map is obtained from the classification of the images of SPOT. We also transformed the raster data to the vector data and found it very efficient and accurate. 2. Image processing and classification: (1)In radiation correction, red and green spectrum was selected as the beginning bands. The difference of the two spectrums was not significant. This approach saved us time and computer memory. (2)Rivers and landslide sites couldn't be classified further. This is due to our study area is located close to the headvaters of the watershed. The channels of river are V-shaped and have high erodibility, narrow width. The spatial resolution of SPOT is only 20Mx20M. Consequently, we couldn't make further classification. However, we could distinguish them from their linear or compact figure with naked eyes. (3)After overlaying the coverage obtained from image classification, the areas with land use changes was detected. The main features were:(A) King-Ment-Tong fault was eroded upward. By analyzing the images, the tendency could be discovered. If the time interval of comparison were long enough the situation may be detected more clearly. (B) During the four-year interval, forest land and grass land have decreased 1.29 and 2.76% respectively, while the area of landslide sites increased 4.05%. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。