無人飛行載具的控制點分布對空拍拼接成果之影響

The influence of Control Point Distribution of Unmanned Aerial Vehicles for the Results of Aerial Stitching

許中立(Chung-Li Hsu) ； 柯典佑(Tien-Yu Ko) ； 袁恆德(Heng-De Yuan) ； 鍾宇泰(Yu-Tai Chung)

《坡地防災學報》 18卷1期 (2019/06) Pp. 1-20

摘要

無人飛行載具是現今航空遙測科技之發展重要方向，且逐漸成為一項三維空間獲取之必備技術，發展至今無人飛行載具飛行穩定度大幅提升，搭配空間定位微型化，使得無人飛行載具於續航時間、影像即時傳輸、高危險地區探測，等各種場合得以充分發揮。以往自然災害發生後仰賴傳統技術取得現地狀況，無法於第一時間取得現場資訊，但應用無人飛行載具之即時性、便利性，掌握救災地表之即刻狀況，為救災相關單位提供實質助益之資料。本研究使用Pix4Dmapper影像處理軟體可快速製作正射影像，期可建構立體視圖與地形資料之三維地形高程模型等等。而將所攝得之影像透過軟體進行拼接，並藉由控制點分布位置不同進而探討其空拍拼接影像之測量精度變化。研究結果顯示，發現控制點數量過多對於精度提升無明顯的幫助，就本研究區域而言選擇20個控制點之均勻分布路線效果最為良好；在越複雜的地形中，控制點應選擇包含各種高程變化的位置上布設；若再無法均勻布設控制點的情況下，建議可以選擇環狀路線方式辦理。

關鍵字

無人飛行載具 ； 地面控制點 ； Pix4Dmapper ； 測量精度

並列摘要

Unmanned aerial vehicles are an important development direction of today's aviation telemetry technology, and gradually become an essential technology for three-dimensional space acquisition. Up to now, the flight stability of unmanned aerial vehicles has been greatly improved, and the space positioning is miniaturized, enabling unmanned aerial vehicles to fully play in various occasions such as battery life, instant image transmission, and detection in high-risk areas. In the past, natural disasters relied on traditional technology to obtain local conditions, and it was impossible to obtain on-site information in the first time. However, the immediacy and convenience of unmanned aerial vehicles were used to grasp the immediate situation of disaster relief sites, providing information for substantial benefits to disaster relief related units. In this study, the Pix4Dmapper image processing software can be used to quickly create orthophotos, and the three-dimensional terrain elevation model of stereoscopic view and terrain data can be constructed. The captured images are spliced through the software, and the measurement accuracy of the aerial stitching images is discussed by the different distribution positions of the control points. The results show that the number of control points is too small to improve the accuracy. For the study area, the optimal distribution route of 20 control points is the best. In more complex terrain, the control points should be selected to be placed at locations containing various elevation changes. If it is no longer possible to evenly arrange the control points, it is recommended to select the circular route method.

並列關鍵字

Unmanned aerial vehicle ； Ground surface control point ； Pix4Dmapper ； Measurement accuracy