| 基于高光谱影像分类线性回归的古画污渍虚拟恢复 |
| |
| 引用本文： | 周平平,侯妙乐,赵学胜,吕书强,胡云岗,张学东,赵恒谦.基于高光谱影像分类线性回归的古画污渍虚拟恢复[J].地理信息世界,2017,24(3). |
| |
| 作者姓名： | 周平平 侯妙乐 赵学胜 吕书强 胡云岗 张学东 赵恒谦 |
| |
| 作者单位： | 1. 中国矿业大学(北京)地球科学与测绘工程学院,北京,100083;2. 北京建筑大学测绘与城市空间信息学院,北京,100044 |
| |
| 摘 要： | 污渍是古画中最常见病害之一,识别并虚拟修复古画污渍是对古画进行科学修复的关键。本文提出了一种基于高光谱影像分类线性回归的虚拟修复方法。首先,利用人工神经网络对受污渍影响较小的波段进行分类,获取油渍区域被遮挡的信息,然后通过建立分类线性回归方程,利用受油渍影响较小的波段去校正受油渍影响较大的真彩色显示波段,在尽量地保持了古画原有的色彩基础上,有效地提取了被油渍遮挡的隐含信息,并一定程度地恢复了画作原貌。
|
| 关 键 词： | 古画 高光谱 分类线性回归 油渍 虚拟修复 |
Virtual Restoration of Ancient Painting Stains Based on Classified Linear Regression of Hyper-spectral Image |
| |
| ZHOU Pingping,HOU Miaole,ZHAO Xuesheng,LV Shuqiang,Hu Yungang,ZHANG Xuedong,ZHAO Hengqian.Virtual Restoration of Ancient Painting Stains Based on Classified Linear Regression of Hyper-spectral Image[J].Geomatics World,2017,24(3). |
| |
| Authors: | ZHOU Pingping HOU Miaole ZHAO Xuesheng LV Shuqiang Hu Yungang ZHANG Xuedong ZHAO Hengqian |
| |
| Abstract: | Stains is one of the most common diseases in the ancient paintings,and the identification and virtual restoration of the painting stain is the key to the scientific repair of old paintings.This paper proposed a virtual restoration method of ancient painting stains based on classified linear regression of hyperspectral image.First,Artificial Neural Network was used to classify the bands affected less by the stains to get covered information in grease region;then classified linear regression model was established to correct the grease much affected true-color bands with the grease less affected bands,which could extract the grease hidden information effectively on the condition of keeping the original color of old paintings,and restored the original appearance of the paintings to some extent. |
| |
| Keywords: | ancient paintings hyperspectral image classified linear regression stains virtual restoration |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
