以模糊四元樹作空間資料分析之研究

　　　　　地理資訊系統是資訊科技應用項目中一項相當重要的成果,對於許多的應用如都市 規劃、交通建設或國防軍事等,均必需仰賴地理資訊系統提供豐富的空間資訊。而四元樹則被 應用於網格式地理資訊系統,作為儲存空間資料所採用的資料結構。由於儲存空間資料通常會 用大量的記憶體,並連帶造成拉長處理時間的嚴重問題。基於四元樹資料結構具有壓縮記憶空 間的優點,故獲得許多研究的重視與採用。本研究除提出藉模糊四元樹架構以編碼方法來表示 大量的二元網格影像資料外,並且利用此模糊四元樹架構來定義許多空間運算子,以顯示模糊 四元樹的應用。因為模糊四元樹架構所需節點數目遠較一般四元樹架構來得少,故以此編碼方 法為基礎,對二元網格影像資料進行空間分析,可以節省大量儲存空間和處理時間。此外,對毋 需非常精確結果卻需具有親和性及快速處理功能的不同地理應用層面而言,此編碼方法頗適 用。
　　　　　Geographical Information Systems (GIS) are generally viewed as significant development of information technology, a bundle of spatial information are provided by GIS to applications such as urban planning, traffic construction or military utilizations, etc. Quadtree is generally applied in raster-based GIS. Quadtree is a data structure to store spatial data. There is a serious problem that the time interval for spatial data processing is enlarged since a huge volume of space is required to store spatial data. Quadtree data structure has been accepted and adopted to alleviate this problem since reduced storage space is found. This report presents not only an encoding scheme to express a large binary raster image by a fuzzy quadtree structure, but also various spatial functions on this structure. Based on this scheme, the space to store an image and the time to perform spatial analyses on this image will be reduced since the image represented by this structure less number of nodes than of an ordinary quadtree structure. In addition, this scheme seems to be well suited for the diverse geographic applications which do not need accurate results but friendly processes with fuzziness and fast processing time.