基于图割高质量深度图获取方法.PDFVIP

Digital video 数字视频 文章编号：1002-8692（2011 ）15-0037-04 基于图割的高质量深度图获取方法 1，2 1，2 1，2 左一帆 ，安 平 ，张兆杨 （1. 上海大学 通信与信息工程学院，上海 200072；2. 新型显示技术及应用集成教育部重点实验室，上海 200072） 【摘 要】3DTV 作为下一代视频广播系统，还有许多技术难点有待解决，其中深度估计是3DTV 的关键技术之一。为了获取高 质量的深度图，提出基于图割（graph cut ）的深度估计方法。该算法在构建能量函数的数据项时，通过对窗口内各个像素赋予 自适应权重，引入梯度信息以抑制因亮度差异导致的误匹配问题并保护边缘信息。然后，经过交叉检测将深度图像素分为可 靠点与不可靠点两类。对检测后的深度图进行后处理迭代优化，从而提高所获取深度值的可靠性。实验表明此算法估计出的 深度图用VSRS 绘制虚拟合成视时比标准的深度估计软件DERS5.1 可有效提高虚拟视质量。 【关键词】多视点；深度图估计；图割 【中图分类号】TN919.8 【文献标识码】A High Quality Depth Map Estimation Method Based on Graph Cut ZUO Yifan1，2, AN Ping1，2, ZHANG Zhaoyang1，2 （1. School of Communication and Information Engineering, Shanghai University, Shanghai 200072, China; 2. Key Laboratory of Advanced Display and System Application of the Ministry of Education, Shanghai 200072, China） 【Abstract 】As the next generation of broadcast system, there are many technical difficulties need to be resolved in 3DTV, in which depth estimation is one of the key techniques. To improve the depth estimation precision, a novel depth estimation method based on graph cut is proposed in this paper. A local window with adaptive support weights is used to compute the average value difference of the pixels ’luminance, which is a part of data item of the cost function. To resist mismatch caused by brightness difference and protect the edge information, gradients between current pixel and its eight neighboring pixels for each pixel in the reference view and current view are computed also. Then, the pixel-wise depth map is cla