计算机工程 ›› 2018, Vol. 44 ›› Issue (7): 237-243.doi: 10.19678/j.issn.1000-3428.0047252

• 图形图像处理 • 上一篇    下一篇

基于马尔可夫随机场的无缝纹理映射

王淼,耿国华,史重阳,余凡   

  1. 西北大学 信息科学与技术学院,西安 710127
  • 收稿日期:2017-05-18 出版日期:2018-07-15 发布日期:2018-07-15
  • 作者简介:王淼(1991—),女,硕士研究生,主研方向为计算机图形学、可视化技术;耿国华(通信作者),教授;史重阳、余凡,硕士研究生。
  • 基金项目:

    国家自然科学基金面上项目(61673319,61373117)。

Seamless Texture Mapping Based on Markov Random Field

WANG Miao,GENG Guohua,SHI Chongyang,YU Fan   

  1. School of Information Science and Technology,Northwest University,Xi’an 710127,China
  • Received:2017-05-18 Online:2018-07-15 Published:2018-07-15

摘要:

在马尔可夫随机场纹理映射技术中,偏向于选择分辨率较低的模糊视图,并且纹理映射存在缝隙。为此,对多角度拍摄产生的图像序列建立三维几何模型,提出一种具有真实感的无缝纹理映射算法。利用能量优化方法为三维模型每个三角面片选择最佳的纹理图像,对其 纹理数据进行精简,节约其存储空间,使用全局一致性颜色调整和局部泊松编辑消除纹理缝隙,得到高质量的无缝纹理映射结果。实验结果表明,与传统的马尔可夫随机场纹理映射算法相比,该算法在胡人俑和大型建筑物上得到了更好的纹理映射效果。

关键词: 多视图纹理映射, 纹理优化, 无缝纹理, 三维重建, 马尔可夫随机场

Abstract:

To solve the problem of view selection with the low resolution and texture mapping with the gap from the technology of texture mapping based on Markov random field,A new algorithm based on image sequences produced by multi angle shooting is proposed for 3D geometric model to achieve seamless texture mapping.Using the method of energy optimization for the 3D model of each triangle to choose the best of the texture image,Simplifying the texture data,saving storage space,using the global consistency of color adjustment and local Poisson editing vanish of texture gap,seamless texture mapping results in high quality.The experimental results show that compared with the traditional Markov random field texture mapping algorithm,this algorithm has better texture mapping effect on Hu people Figurines and large buildings.

Key words: multi-view texture mapping, texture optimization, seamless texture, 3D reconstruction, Markov random field

中图分类号: