多重亮度紅外線打光器之影片摘要

Abstract

在夜間影像監控系統中，適當的照明對於影像品質扮演相當重要的角色。就一般紅外線打光器而言，因受限於亮度固定，可能因亮度不足造成遠距離的物體難以辨識；而光源充足的情形下，過於接近攝影機的物體則會出現影像過曝，以至於前景影像品質低落。 在本篇論文中，我們針對新型多重亮度紅外線打光器產生的不同亮度階層之影像進行影片摘要。首先，我們利用高斯混和模型對各個亮度階層的影像進行前景擷取，並透過影像品質評估，使系統能自動於各亮度中選擇出最符合人眼視覺的前景，而形成新的影像輸入。最後，我們設計出自動影像摘要方法以選出輸入影片中具代表性的關鍵影像，並將關鍵影像與事先選好的背景合併產生摘要影片。

In nighttime video surveillance, proper illumination plays a key role for the image quality. For ordinary IR-illuminators with fixed intensity, faraway objects are often hard to be identified due to insufficient illumination while nearby objects may suffer from over-exposure, resulting in image foreground/background of poor quality. In this thesis we proposed a novel video summarization method which utilizes a multi-intensity IR-illuminator to generate images of human activities with different illumination levels. First, a GMM-based foreground extraction procedure is adopted for images acquired under each illumination level. With quality assessment of the outcome of such procedure, the system than selects visually most plausible foreground regions from different illumination levels to generate a set of new input data. Finally, an automatic video summary method is developed to identify key frames for these data and merge them with a preselected representation for still background. The result brings out a reasonable video summary for moving foreground, which is generally unachievable for nighttime surveillance videos.