摘要
对动态场景进行光场渲染时,场景中的运动物体通常需要满足线性、低速运动的条件,并要求相机具有较高帧率。本文提出了一种新的对动态场景的光场渲染方法。该方法通过获得运动物体在不同位置处的信息角,对场景中非线性运动进行适当的转化,实现对动态场景的多视点渲染。该研究结合并推广了静态光场渲染的方法,放宽了对物体运动的要求,并适用于场景中非线性运动的物体。和现有的方法相比,该方法具有计算和存储效率高、适应性强、算法鲁棒性强的优点。实验结果进一步体现了该方法较好的应用前景与可操作性。
For light field rendering of dynamic scenes, it often requires that the moving obJeCt should move nnearty wnn low speed, and it also requires that the cameras have high frame rates. This paper proposes a novel method to render dynamic scene which expands the aspects of light field research. If we know the light field rendering information of the object in its original position at^d information angle, meanwhile the object doesn't have great transmutation during its movement, we could gain light field information in any new position using this method. The conversion reduce frame~ rate requirement of cameras and algorithmic complexity, and realize the purpose of rendering dynamic scene through multi viewpoints. Compared with the existing method, this method has many advantages such as high storage efficiency, adjustability, and good robustness. Experimental results show that our method is capable of rendering nonlinear moving objects with photorealistic quality.
出处
《电子测量技术》
2009年第9期55-59,共5页
Electronic Measurement Technology
关键词
信息角
动态场景
光场渲染
线性运动
转化
information angle
dynamic scene
light field rendering
linear movements
conversion