Journal of Systems Engineering and Electronics

• 软件、算法与仿真 • 上一篇    下一篇

基于Mean Shift 的光晕抑制算法

赵蓝飞,席志红   

  1. 哈尔滨工程大学信息与通信工程学院,黑龙江 哈尔滨 150001
  • 收稿日期:2012-05-10 修回日期:2013-03-26 出版日期:2013-07-22 发布日期:2013-05-15

Halo inhibition algorithm based on Mean Shift filter

ZHAO Lan-fei,XI Zhi-hong   

  1. College of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, China
  • Received:2012-05-10 Revised:2013-03-26 Online:2013-07-22 Published:2013-05-15

摘要:

针对SSR 算法在图像高对比度边缘处产生光晕现象,在研究光晕现象的基础上,提出了一种基于均值偏移(Mean Shift,MS)的自适应滤波光照估计算法。该算法根据初次MS 算法平滑原图像得到的先验知识,自适应地改变滤波器的形状,从而达到抑制光晕的目的。实验表明该算法不仅压缩了图像的动态范围,而且运算速度较快,抑制光晕效果较好。

Abstract:

Due to SSR algorithm is suffer from halo artifacts which are introduced by the high range dynamic (HDR) edge, this paper makes a research on the halo artifacts and propose an illumination estimation algorithm which can inhibit halo artifacts, based on Mean Shift (MS) and self-adaptive filter. This algorithm alters the shape of the self-adaptive filter according to the prior knowledge obtained from the image which is smooth by the first time MS filter, thus halo artifacts are restrained. Experimental results demonstrate that this algorithm makes range dynamic compression, has a faster calculation speed and a better effect on halo artifacts inhibition.