為因應未來三維地理資訊之應用需求，三維製圖為重要之工作。本產學合作研發案將延伸產業界之航測製圖功能，開發三維航測製圖之全方案。本計畫以三年為期建立三維製圖之方案，計畫重點為配合標準之航測製程建立三維地理資訊所需之房屋及道路模型，並套合影像及模型產生真實正射影像，最後以立體瀏覽方式進行動態展示及查詢。第一年之研發內容為連結傳統航測立體量測之三維線段產生三維房屋及道路模型。研究重點包含房屋結構線二維位相關係重建及三維形塑，道路邊界線之重組及路面重建。第二年之研發內容為使用重建之三維模型與航照影像進行套合，以產生真實正射影像及房屋屋頂之紋理。研究重點為遮蔽區偵測與補償，及多幅影像之色彩平衡與鑲嵌。第三年，即本年，之研發內容為由二維圖資建構三維模型之流程，並建立一個可操作之三維瀏覽系統，供三維動態瀏覽、屬性查詢。研究重點為三維模型之建置及多層次細緻程度處理、動態瀏覽顯示、應用分析功能與圖形化使用者界面等技術之研發。全程研發工作包含方法之建立、系統雛型之建置、測試及驗證。 To meet the demand from 3-D geographic information systems, 3-D mapping is an important work. This project will extend the standardized procedure in photogrammetric mapping to establish a multi-function system. The highlight of this three-year project is to enhance the current production procedure then to generate 3-D building and road models. A registration between aerial photographs and the generated models is also performed for the derivation of true orthoimages. A 3-D browse system will also be developed for 3-D representation, query, and application analyses. The work in the first year links with the photogrammetric mapping system to acquire 3-D building/road lines for the reconstruction of models. The core task includes the 2-D topology building and 3-D shaping. The work in the second year performed image-model registration to generate true orthoimages. The roof image may, thus, be derived for roof texture mapping. The major works are occlusion detection and compensation, mosaicking and color balancing for multi-frame images. In the third year, the primary work is to develop algorithms for constructing 3D building models from existing 2D CAD data and to implement an operable system for 3-D browsing, query, and application analyses. The technique for treating multi-level of details will also be developed in the third year. The research and developed work include methodology, establishment of a prototype, testing and validation. 研究期間：10002 ~ 10101
