深度图由锐边缘包围纹理的匀质区域组成,因此边缘信息对于虚拟视的绘制至关重要.由于深度图不同区域的失真会对绘制虚拟视质量产生不同的影响,提出一种基于边缘的面向虚拟视绘制的深度编码算法.将深度图划分为边缘区域和平坦区域:对于边缘区域,采用全搜索策略,以提高边缘区域的编码质量;对于平坦区域,采用SKIP模式和16×16模式,以提高编码效率.除此之外,在去块滤波中,提出一种利用纹理图和深度图边缘相似性的可保护深度图片边缘的中值三边滤波器.实验结果表明,与JM编码方案相比,该方法在码率基本不变的前提下,提高了编码效率和虚拟视绘制的主观质量.

The depth information of the scene indicates the distance between the object and the camera, and depth extraction is a key technology in 3D video system. The emergence of Kinect makes the high resolution depth map capturing possible. However, the depth map captured by Kinect can not be directly used due to the existing holes and noises, which needs to be repaired. We propose a texture combined inpainting algorithm in this paper. Firstly, the foreground is segmented combined with the color characteristics of the texture image to repair the foreground of the depth map. Secondly, region growing is used to determine the match region of the hole in the depth map, and to accurately position the match region according to the texture information. Then the match region is weighted to fill the hole. Finally, a Gaussian filter is used to remove the noise in the depth map. Experimental results show that the proposed method can effectively repair the holes existing in the original depth map and get an accurate and smooth depth map, which can be used to render a virtual image with good quality.

Multi-view images captured by a parallel multi-camera array manually placed tend to produce vertical parallax. To eliminate vertical parallax, the multi-view images must be rectified. This paper presents a simple and effective algorithm for multi-view rectification, and builds a stereoscopic display system based on this algorithm. A multi-thread technique is used to achieve soft synchronization of the cameras, and to perform real-time optimization. When image resolution is 1 280×960, the upload rate is 25 frame/s, and the target 3D display resolution is 1 920×1 080, the display rate of the system can reach 20 frame/s by using the bilinear interpolation. If interpolation methods with low complexity are used, the display rate can reach more than 20 frame/s. A 3DTV system with real-time rectification of eight cameras and stereoscopic display is built by using the proposed algorithm.

在多视自由视点电视(FTV)系统中,摄像机阵列拍摄的多视视频存在垂直视差、且水平视差分布不均等问题,严重影响合成视频的质量效果.现有的多视图像校正方法虽然消除了垂直视差,但是校正图像过程中引入的旋转和平移失真较大,使人眼产生不舒适感.为此,本文提出一种适于平行摄像机阵列的多视图像校正算法,首先建立理想的摄像机安放位置,然后计算多视图像校正变换,将实拍图像映射到理想摄像机位置处,得到校正后图像.在设计理想的摄像机安放位置时,考虑人眼视觉特性,确保坐标轴旋转和平移调整量最小化,实现多视图像校正变换引入图像失真总量最小.实验表明,该算法不仅可以消除垂直视差,而且图像校正过程中引入的失真小,符合人眼视觉习惯,能有效改善合成视的图像质量,提高人眼观看舒适程度.

The quality of the virtual viewpoint determines the effect of the 3D video on the display terminal.To fill the holes generated in the depth image-based rendering algorithm,we propose the dual texture + dual depth multi-view rendering algorithm in this paper.The left reference texture image and its corresponding depth image are used to render the virtual image by applying DIBR technology,and to draw the mask image from the filtered virtual image.Then each point coordinate of large holes in the mask image is inversely transformed to the corresponding pixel coordinate in the right reference texture image to get the hole-filling image of the virtual image based on the judgment of whether the depth value of the pixel belongs to the background.Finally,fuse the hole-filling image with the virtual image obtained through DIBR from the left view to fill the large holes,and fill the small holes by using the interpolation algorithm.Experiments show that our method can effectively inpaint the big holes due to DIBR,so that virtual images with good quality can be got.