文章詳目資料

國防管理學報

  • 加入收藏
  • 下載文章
篇名 IKONOS影像萃取DEM提升幾何精度之研究
卷期 26:2
並列篇名 A Study on Extracting Digital Elevation Models Refine Geometric Corrections Accuracy by Using IKONOS Imagery
作者 張學賢呂鴻道傅振華
頁次 037-052
關鍵字 數値高程模型幾何糾正正射化糾正影像匹配Digital Elevation ModelGeorectificationOrthorectificationImagery Registration
出刊日期 200511

中文摘要

近年來運用傳統航空攝影來獲取地表三維幾何資訊,雖已漸趨成熟,但欲獲取高山 或無法量測地區之幾何資訊仍有許多不足之處,自從IKONOS-2衛星於1999年9月成 功升空,可拍攝高解析力、高穩定度與高精度的影像,對於需使用高精度幾何資訊的需 求,獲得了具體的改善,這也使衛星遙測影像與處理技術運用進入一嶄新的世代。 衛星影像因受傾斜攝影與地形起伏等幾何變形影響,因此使用前需先用地面控制點 進行幾何糾正(Georectification)或數位高程模型做正射化糾正(Orthorectification),使影像 達到一定精度(如製圖精度),才可做後續許多加値運用。在高山地區或不可測地區無法 進行地面控制點量測,因此本研究嘗試以嘉義地區單幅、立體像對影像,分別在有、無 地面控制點之下,進行有理數函數模之幾何糾正與運用影像套合(Imagery Registration)技 術萃取數値高程模型,進行嚴密模式之正射化糾正實驗,並觀察其均方根誤差値(Root Mean Square Error; RMSE)狀況,分析其是否可提昇影像幾何精度0 综合實驗結果顯示,在無控制點下,僅使用RPC與萃取之DEM執行幾何糾正與正 射化糾正,證實仍可大幅提昇影像幾何精度,這些方法可提供應用於高山地形起伏大或 不可測量地區之量測,以大幅節省時間、成本及維護測量人員的安全。

英文摘要

Although using traditional air photographing to get the ground information of 3D geometrical has become a mature technique in recent years. But it is still not satisfactory in the high mountain and unidentified areas. However, since the success launch of IKONOS 2 on 24 Sep. 1999; high resolution, stable and accurate image of IKONOS 2 made a great improvement to fit the request of high accuracy geometrical information. This also brings the satellite remote sensing image and the processing technique into a new generation. The satellite image is affected by the geometrical factors such as inclined photography and the undulation of the ground shape. For reaching a certain accuracy (exp. Mapping accuracy), it must use the ground control points to go through the georectification or use DEM to go through the orthorectification before producing the product and then make various additional applications. In the high mountain and unidentified areas, the ground control points can not be estimated. Therefore, this research is trying to use single and stereo pairs image of Gia-I(^ A) area separately with control point and no control point to observe its Root Mean Square Error, RMSE and also make it to go through the georectification of the rational function, Imagery Registration, extract the DEM and orthorectification experiment to analyse if it could improve the geometrical accuracy of the imagery. The results of the comprehensive experiment show that the orthorectification and georectification execution with RPC and DEM extraction also can improve geometrical accuracy of the imagery very much. These methods could be used to estimate the mountain and unidentified areas in order save time, prime cost and also keep the security of the estimation personnel.

相關文獻