上海交通大学学报

上海交通大学学报（自然版） ›› 2011, Vol. 45 ›› Issue (10): 1509-1515.

• 自动化技术、计算机技术 • 上一篇    下一篇

基于Contourlet的改进加权抛物线插值图像超分辨率算法

 蔡念, 张海员, 张楠   

  1. (广东工业大学 信息工程学院， 广州 510006)
  • 收稿日期:2010-10-08 出版日期:2011-10-31 发布日期:2011-10-31
  • 基金资助:

    国家重点基础研究发展规划（973）项目（513230202032），国家自然科学基金（61001179，30600121），广东省自然科学基金（07301038）资助项目

Using Improved Weighted Parabolic Interpolation and Contourlet Transformation to Zoom Images for Super-resolution

 CAI  Nian, ZHANG  Hai-Yuan, ZHANG  Nan   

  1. (School of Information Engineering, Guangdong University of Technology, Guangzhou 510006, China)
  • Received:2010-10-08 Online:2011-10-31 Published:2011-10-31

摘要： 为了尽可能地保持图像的基本信息，提高图像的视觉效果和空间分辨率，提出一种基于Contourlet的改进加权抛物线插值算法.该算法增加加权抛物线插值的误差补偿项，利用Sobel算子设定插值点的边缘方向，得到初始放大图像.利用Contourlet提取高频成份，原始图像幅值增强充当低频部分，再经过Contourlet逆变换得到高分辨率图像.实验结果表明，相对于传统的图像放大算法，该算法考虑了全局相关性，得到了更加清晰的边缘信息.

关键词: 图像插值, 图像放大, 超分辨, Contourlet变换, 误差补偿

Abstract:  In order to preserve the basic information of images and to improve their visual effect and space resolution, a novel image zooming algorithm was proposed by using improved weighted parabolic interpolation and Contourlet transformation. The proposed algorithm incorporates the erroramended part into the weighted parabolic interpolation algorithm. Directions of the interpolation points are determined by the Sobel operator, then we get a preliminary zoomed image. Highfrequency components are got by a Contourlet transformation, and lowfrequency components are replaced by the enhanced amplitude of the original image. A highresolution image is achieved by inverse Contourlet transformation. The experimental results indicate that, compared with traditional image zooming algorithms, this algorithm can get clearer and sharper edges due to considerations of global pixels correlation of the original image.

Key words:  image interpolation, image zooming, superresolution, Contourlet transformation, error compensation

中图分类号: