臺灣博碩士論文加值系統

　　地理資訊系統的功用在於整合與管理空間資料與其屬性，以提供地理相關的資訊，作決策與分析之用。隨著電腦軟體技術的進步，地理資訊系統的應用需求也日益突顯。除了地政管理、都市土地使用規劃、林地管理、公共管線管理等應用外，在軍事上則可作為戰術、戰略決策上的支援與應用。灰階及彩色地圖、影像資料佔用大量儲存空間，一直是地理資訊系統極需克服的問題，本文提出一新的線性四元樹編碼法，稱為位元分離線性四元樹編碼法（bit separation linear quadtree coding scheme, BSLQ）。其運用在地理資訊系統中灰階及彩色地圖、影像資料儲存的問題上，可將編碼時間維持在與圖形解析度成線性親係之條件下，並可降低儲存灰階及彩色地圖、影像資料所需之記憶體空間。此外,文中所提之編碼結果，因具有階層關係及四元樹分割的特性，故可直接應用在邊緣偵測之運算，以達到降低資料運算量的效益。在四元樹的單一像元階層之終端節佔所表示的影像資料即有影像的部份邊緣資料，所以邊緣偵測只需對四元樹中多像元階層的終端節點做處理處理，即可找出影像資料的實際邊緣。由於是對本編碼結果直接處理，且只需對四元樹中多像元階層的終端節點做處理運算，故可減少邊緣偵測的資料處理量。另外，為驗證與測試所提之編碼法及邊緣偵測方法，本文中以六張灰階及四張彩色影像進行實證比較，結果亦顯示所提方法之優越性及可行性。

　　A geographical information system (GIS) is usually applied to integrate and manage spatial data with associated attributes. The purpose is to provide geographical information to support decision making and analysis. The importance of the GIS is even more significant as the rapid development of computer software technology is found. In addition to the usage of GIS in land policy management, municipal land use planning, public pipeline management, the GIS is also applicable in military. Strategic and tactic decisions are believed to be closely related to GIS. When the gray level or color images are used in a GIS, a large volume of space is used to store images. The requirement of a large storage space renders serious problems in data processing or transmission. A linear quadtree coding scheme, bit separation linear quadtree coding scheme (BSLQO, is proposed in this paper, The effect of the proposed BSLQ scheme can alleviate the storage problem by using a method with linear complexities of time and space to construct a quadtree. Operation of edge detection is also provided by fully making use of the hierarchical structure of quadtrees. The data structure of linear quadtrees is appropriate to detect edge since the last layer of a quadtree consists of only single-pixel level data. An edge can be easily detected if the leave nodes with associated neighbor pixels are searched and connected. Several empirical tests are implemented to analyze the characteristics of the proposed coding scheme. Six gray level images and four color images are trailed to derive the results. All results reveal that not only the compression effect of the proposed scheme can be affirmed, but also the edges in testing images can be completely found. The results of empirical tests approve the feasibility of the proposed scheme.