欢迎访问《图学学报》 分享到:

图学学报

• 视觉与图像 • 上一篇    下一篇

基于阈上随机共振的侯马盟书虚拟修复方法

  

  1. 中北大学计算机与控制工程学院,山西 太原 030051
  • 出版日期:2017-06-30 发布日期:2017-07-06
  • 基金资助:
    山西省重点研发(指南)项目(201603D321128)

Virtual Restoration Method of Houma Convenant Tablets Based on Suprathreshold Stochastic Resonance

  1. School of Computer and Control Engineering, North University of China, Taiyuan Shanxi 030051, China
  • Online:2017-06-30 Published:2017-07-06

摘要: 针对山西侯马盟书严重退化问题, 对其文字的提取和虚拟修复问题进行了研究。将
图像处理算法与人机交互相结合,提出一种基于阈上随机共振的侯马盟书虚拟修复方案。首先将
原始RGB 盟书图像变换到色调、饱和度、亮度(HSV)颜色空间,再利用阈上随机共振对亮度图像
进行复原,然后通过改进的双边滤波Canny 算子提取文字轮廓信息,并结合人机交互对确实的文
字轮廓进行修复;最后利用颜色迁移算法将经过填充的文字恢复成原有朱砂红颜色。实验结果证
明,该方法能有效修复笔画连通性差的模糊字迹和褪掉的色彩。

关键词: 侯马盟书, 虚拟修复, 随机共振, 轮廓提取

Abstract: On the issue that Shanxi Houma convenant tablets is serious degraded, this paper does
researches on extraction the text and virtual restoration. Combining image processing algorithms and
human-computer interaction, a method of Houma convenant tablets virtual restoration based on
suprathreshold stochastic resonance was proposed. Firstly, transform the original RGB image to hue
saturation value (HSV) color space, and restore the brightness image by suprathreshold stochastic
resonance (SSR), then use modified bilateral filtering Canny operator to extract text contour
information, and combine with human-computer interaction to text contour restoring, finally, adopt
the method of color transfer to restore the text after filling into the original cinnabar color.
Experimental results show that the fuzzy words with poor stroke connectivity and faded color of
Houma convenant tablets can be repaired effectively by digital image processing.

Key words: Houma convenant tablets, virtual restoration, suprathreshold stochastic resonance;
contour extraction