本研究之目的在於建立一個機器影像視覺系統，將無人飛行器於各個不同視角所空拍得之影像畫面整合於同一平面，使之能用於後續的整體分析應用，並兼負有目標物定位與導航應用的功能。
當無人飛機配掛有影像酬載時，本系統可對影像攝影機所拍得的畫面進行鏡頭扭曲的徑向校正，並將三度空間所攝得的影像配合載具的姿態參數進行影像的旋轉、正射化轉換，而後再後利用GPS位置在二維的平面上將之進行整合。如此一來不僅可對畫面中所攝得地面上的任一點進行目標物的定位，也可將轉換整合後所得之影像地圖用於航機的導航。
整個研究的理論方法及全套系統皆已在實際飛行測試中獲得驗證。透過飛機所即時下傳的影像及姿態、位置參數，可立即於畫面中選擇目標點並成功進行定位，也可將影像經正射化後整合成一幅區域的影像地圖，又可透過後處理的方式將更高精度的飛行資料與錄製的影像進行計算，得到更理想的準確度。

The purpose of this thesis is to construct a vision-based system which can combine the images taken via UAV from different orientations into a single plane, so that the processed images can be analyzed and applied to navigation and target positioning applications.
When the UAV has an imaging system, after we obtain the 3-D images along with the attitude parameters, the images are rotated and ortho-image are created. After that, GPS position fix on 2-D plane is integrated with the processed images. Therefore, we can not only locate the ground target’s coordinate which is in the camera’s view, but also utilize the resultant images which can be served as navigation map.
The whole system and the methodology been verified via flight test. Through the real time down-linking of attitude, position, and images from UAV, we can pinpoint a target and displayed those aerial photos on the ground station’s monitor in real time successfully. Besides, it also maps the ortho-images into an local area image map. By introducing the post-processing method, high accuracy flight data from onboard computer, along with the recorded images can result in a more accurate solution.

[1] Optical sensing payload -http://www.noaanews.noaa.gov/stories2005/s2421.htm
[2] RMRL website -http://rpv.iaa.ncku.edu.tw/
[3] Y.N. Liu, “Vision-based Moving Target Tracking and Estimation for Unmanned Helicopter”, Master thesis, Institute of Aeronautics and Astronautics, National Cheng Kung University, June 2009.
[4] P.Y. Huang, “Realization of Vision-based Runway Detection for Fixed-wing UAV”, Master thesis, Institute of Aeronautics and Astronautics, National Cheng Kung University, June 2009.
[5] F. Caballero, L. Merino, J. Ferruz, A. Ollero, “Unmanned Aerial Vehicle Localization Based on Monocular Vision and Online Mosaicking”, J. Intell Robot. Syst, January 2009.
[6] 鍾國亮,“影像處理與電腦視覺導論”,台灣東華出版,台北市,中華民國九十七年
[7] Wikipedia - http://www.wikipedia.org/
[8] C.W. Cheng, “Projective Geometry in Photogrammetry”, Master thesis, Department of Civil Engineering, National Taiwan University, 2007.
[9] WGS84 to TM2-97 transformation –
http://gis.cy1000.com.tw/transcoor/transcoor.asp
[10] Google Earth - http://earth.google.com/
[11] Google Map - http://maps.google.com.tw/