一种基于深度图的三维/多视点视频视点合成方法

摘要

基于深度图的视点合成技术已成为国际视频技术研究的热点之一.本文实现的基于深度图的三维/多视点视频视点合成算法,在点对点的三维变换的基础上增加了深度图预处理、去像素重叠、基于空间对象的插值等优化,此外试图针对不同的应用场合,对变换图像进行不同的空洞填补方法.实验结果表明,本方法最后合成的视点图像具有良好的视觉效果.%View synthesis has become one of the hot spots of international video technology. The realization of a 3D/multiview video view synthesis method based on DIBR is discussed. The depth map preprocessing, eliminating pixel overlap and spatial object-based interpolation based on the pixel-to-pixel 3D warping are increased.Besides, different ways of hole-filing for different application are trying to utilize. The experiment results show that the method of view synthesis can obtain a satisfactory visual effect.