• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2014, Vol. 36 ›› Issue (11): 2191-2195.

• 论文 • 上一篇    下一篇

基于尺度空间的小波纹理描述算法的壁画修复

任小康,邓琳凯   

  1. (西北师范大学计算机科学与工程学院,甘肃 兰州 730070)
  • 收稿日期:2013-04-07 修回日期:2013-07-04 出版日期:2014-11-25 发布日期:2014-11-25

Murals inpainting of the wavelet texture description
algorithm based on scale space           

REN Xiaokang,DENG Linkai   

  1. (College of Computer Science and Engineering,Northwest Normal University,Lanzhou 730070,China)
  • Received:2013-04-07 Revised:2013-07-04 Online:2014-11-25 Published:2014-11-25

摘要:

近年来,壁画的数字化修复技术得到了很大的发展,但是其主要算法在图像纹理方向的修复效果一般。为了解决这些问题,还原壁画的原有样貌,提出了一种简单快速的壁画修复算法。该算法在改进的Telea算法的基础上,根据图像纹理的多尺度特征性,利用小波纹理描述算法对壁画的纹理结构进行了修复处理,运用非局部梯度算子,通过计算待修复区域的像素点的非局部梯度值并依据曲率选取的全局最优方向进行扩散,达到壁画纹理修复的目的。实验结果表明了该算法的有效性。

关键词: 壁画修复, Telea算法, 尺度空间, 小波纹理描述算法, 梯度算子

Abstract:

In recent years,murals digital restoration is greatly developed,but its main algorithms has no obvious effect on the restoration of image texture direction.In order to solve this problem and restore the original appearance of murals, we propose a simple and rapid mural repairing algorithm in which the Telea algorithm is improved. According to the multiscale feature of image texture,the proposal repairs texture structure of murals by using the wavelet texture description algorithm. To repair the murals’ texture, nonlocal gradient operator is used to calculate the nonlocal gradient value of pixels in unrepaired areas and diffusion is performed according to the global optimal direction. The experimental results demonstrate the effectiveness of the algorithm. 

Key words: mural inpainting;Telea algorithm;scale space;wavelet texture description algorithm;gradient operator