摘要
为在平面幕上建立立体视觉,通过模拟真实世界中人眼获取图像过程,介绍计算机绘制双眼视差图像的方法。给出虚拟场景中2种放置双眼视见体的模型,通过投影分析可以得出内束模型会引入垂直视差,且水平视差不恒定,从而引起同一深度面的物体被感知深度变形。离轴模型符合人眼建立立体视觉的习惯,可以正确反映出物理屏幕前后不同深度的物体,且比单眼图像的视野范围大。
In order to establish stereo vision on the flat screen,by simulating human eyes＇ process of acquiring images in the real-world,this paper introduces methods of rendering binocular parallax images by the computer.Two models of view frustum placement in the virtual environment are given.Through projective analyzing result can be concluded that toe-in model causes vertical parallax and variance of horizontal parallax which leads to deformation of depth perception in the same depth of virtual environment.Off-axis model conforms to the human eye habit of establishing stereo vision and correctly reflects the depth of the object before and after physical screen.Binocular images created by off-axis model have lager field of view.
出处
《计算机工程》
CAS
CSCD
北大核心
2011年第17期271-273,共3页
Computer Engineering
基金
国家"973"计划基金资助项目(2009CB320805)
关键词
立体显示
视差
深度感知
内束模型
离轴模型
stereo display
parallax
depth perception
toe-in model
off-axis model
