傳統監控系統通常於監控場景旁架設固定攝影機，針對其取得之影像進行分析; 由於背景固
定，故可使用如背景相減或連續畫面差異等技術將背景消除並偵測出移動前景。然而本研究將攝
影機架設於空中飛行之移動載具上，因此要面對的問題包括 : 影像背景持續變動、易受氣流擾亂
導致畫面抖動而增加分析前景困難、移動物體相對於廣闊區域較小且移動速度相對變慢等這些問
題。本研究提出一用於連續空照影像之移動物體偵測方法，藉由影像對齊技術補償飛行器之移動
動量，再藉由計算幀差影像來得到屬於移動物體的像素。另一方面，再套用影像分割的技術，將
影像中具有同質性的像素予以群化，進而融合幀差影像與分割的結果，以漸進式形變方式，取得
合理的移動物體輪廓。最後再藉由物件追蹤的技術，給予移動物體一致性標記並剔除假警報。實
驗結果顯示，針對於空照影像上之物件偵測及追蹤，本研究所提出之混合式演算法表現優於幀差
法及光流法。

Compared with stationary surveillance cameras, moving object detection on a camera carried by an airborne vehicle is more difficult because of the relatively dynamic background. In this study, we present a hybrid method for moving object detection in aerial videos. The ego motion of airborne vehicle is compensated by feature-point based image alignment on the consecutive frames. Then an accumulative frame differencing method is used to detect the motion pixels. Meanwhile, the current frame is divided into homogenous regions by image segmentation, and some of them were selected as candidates of moving objects by prior rules. The motion pixels and the moving object candidates are then fused by a
morphing-based approach to retrieve the information of position and shape of moving objects. Moreover, a Kalman-filter tracker is adopted to give a consistent label on each detected moving object and avoid the false
alarms. The proposed method is evaluated based on the video data captured on an airborne vehicle at different altitude. Experimental results revealed that the proposed hybrid method has better performance than
those frame-difference and optical-flow based approaches.