分析了当前三维视频源获得的方法,指出了从二维视频图像转换成三维视频的现实意义;并从二维图像的视频源特点出发,提出了综合应用颜色深度信息估计、运动深度信息估计和几何透视深度信息估计这三种景深线索的算法,来获得二维图像的景深图,比现有的单一线索估计图像景深有着明显优势。

The method of obtaining the source of three-dimensional video is analyzed; the significance of the transformation from two-dimensional video to three-dimensional video is given. On the basis of two-dimensional video’s characteristic, the algorithm combined color depth estimation, motion depth estimation and geometry depth estimation is proposed, which is used to obtain two-dimensional depth-image. It is better than single cue estimating depth-image.