卷期 : 47 / 3
出版年 : 2019/08/01
在二十一世紀的今日,隨著遙感與控制技術進步及影像酬載性能輕巧化,UAV的應用已逐漸成為新興主流。相較於衛星攝影及航空攝影,UAV飛行高度低的特色可獲取高解析度影像,並進行空間資訊的快速提取,非常適用於環境監測、及時救災、地理測繪等領域。只是,飛行高度低也造成單幅影像所涵蓋的範圍偏小,需拍攝多張影像方能完成大範圍地區攝影,如何在極短時間內透過拼接技術將影像拼接為大範圍影像,已是一項重要研究課題。目前常見的處理方式大多以Agisoft、PhotoScan、ERDAS Imagine、Pix4D Mapper等市售商業軟體為主,於拍攝任務結束後,將影像全數載入進行影像拼接,若輔以地面參考點的座標修正,可在短時間內快速取得超高精度的地區正射影像(地面解析度可達5cm等級)與數值地形模型。隨著UAV各型載具與各類影像攝影技術的應用面向日益廣泛,本文主要針對航空攝影與影像拼接的技術要點與操作經驗作一描述,以供院內同仁對於UAV影像應用的技術認知。 With increasingly advanced technology, high resolution images, and its value-added application, UAV has been playing roles in various missions, such as monitoring environmental change, and disaster photography. However, due to the field-of-view (FOV) limitations, it is required to take more UAV images which need further alignment and stitching. To this end, the current prevailing solutions are using retail software, such as Agisoft, PhotoScan, ERDAS IMAGINE, and Pix4D Mapper. After completing photography missions, UAV images are loaded and transmitted to the processing program for stitching related to geometric correction. Using the coordinate correction with the ground control points (GCP), we can quickly complete panoramic images, obtain high-level digital terrain model (DTM), and get digital orthophoto map (DOM) to the level that the GSD can achieve at 5 cm/pixel. This paper presents a preliminary study on diverse scene stitching technology for aerial video surveillance and the additional benefits for UAV images.
關鍵詞 : 無人載具(UAV)、影像拼接(image stitching)、航空攝影(aerial video surveillance)