應用幾何校正法輔助無人載具空拍影像從事調查分析之研究

Abstract

RPV has been widely used in different missions in recent years. For example, surveying on the development of city infrastructure, researching agriculture, fishery and farming, monitoring on environmental protection, highway driving, forestry management and disaster prevention like floods or mud-rock flows, are the works used by RPV. However, RPV is restricted to congenital conditions and acquiring data is used for differentiation only. In order to increase the accuracy of the images and to concern their function, some corrections are necessary. This study compares several Geometric correction models of Erdas Imagine8.4 to increase practicability of RPV's photo. The result shows that the Rubber Sheeting Method and the Polynomial Method are the two proper models used in correcting images taken by RPV. It also shows that the Non-linear Rubber Sheeting Model is more perfect, and the accuracy is raised with increasing the number of Ground Control Points. The images are successfully simulated in 3-D systems to extending the practical application of RPV. We believe that it can be effectively used to survey and analyze disaster area with increasing the quality of image.

無人載具直昇機為近年來所廣泛使用之遙控飛行器，其所能擔任之任務非常多，舉凡都市發展建設之探勘、農漁牧之調查、環保監測、公路偵搜、森林經營、坡地防災監測、水土災害與海岸偵防之監控、災前監測與災後調查等。然因無人載具直昇機從事航拍作業經常受其先天條件所限，以致所獲取的資料僅能做為判釋之用，必須從事幾何校正以提高影像之精確度，方能充分發揮其功能。 為提昇無人載具直昇機航拍影像之實用性，本研究使用Erdas Imagine8.4軟體就其所有多種不同的幾何校正模組中進行比較，結果以橡皮伸張法（Rubber sheeting）及多項式法（Polynomial）兩種較適用於無人載具空拍影像之幾何校正。經研究得知非線性橡皮伸張法較適用於無人飛行載具直昇機空拍影像之校正，此法可隨控制點數的增加而提高精度。 研究中並將影像成功地套入立體模擬查詢系統，以擴展無人載具直昇機空拍影像之應用方向，除了能做為影像判釋之外，隨著未來影像品質之提昇，相信亦能有效輔助災區調查分析之工作。