计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (24): 177-178.DOI: 10.3778/j.issn.1002-8331.2010.24.053

• 图形、图像、模式识别 • 上一篇    下一篇

一种盲Poisson MAP超分辨方法

杨学峰,李金宗   

  1. 哈尔滨工业大学 图像信息技术与工程研究所,哈尔滨 150001
  • 收稿日期:2009-02-13 修回日期:2009-04-15 出版日期:2010-08-21 发布日期:2010-08-21
  • 通讯作者: 杨学峰

Blind Poisson MAP super-resolution method

YANG Xue-feng,LI Jin-zong   

  1. Institute of Image Information Technology and Engineering,Harbin Institute of Technology,Harbin 150001,China
  • Received:2009-02-13 Revised:2009-04-15 Online:2010-08-21 Published:2010-08-21
  • Contact: YANG Xue-feng

摘要: 由于成像系统内部和外部的因素,使得获取的图像产生退化。为获取高分辨率图像,图像超分辨技术获得迅速发展。基于各向异性正则化提出了一种盲Poisson MAP算法,该算法同时估计高分辨图像和包含了位移因素的模糊函数,与基本的盲Poisson MAP算法相比,有效地抑制了振铃效应,清晰度有很大改善,峰值信噪比大约提高7 dB。

Abstract: The image obtained by some sensors will be degraded unavoidably due to some inner and outer reasons of imaging system.In order to obtain high resolution image,the image super-resolution technology has been developed rapidly.A novel blind Poisson MAP algorithm is established,which may synchronously estimate high-resolution image and blur function containing shift factors.Compared with fundamental blind Poisson MAP algorithm,this algorithm can suppress oscillatory artifacts effectively and improve the definition greatly and increase peak SNR about 7 dB.

中图分类号: