Journal of Applied Sciences ›› 2011, Vol. 29 ›› Issue (2): 159-168.doi: 10.3969/j.issn.0255-8297.2011.02.009

• Signal and Information Processing • Previous Articles     Next Articles

Fusion Strategy of Multi-scale Image Decomposition Combined with Image Quality Evaluation

PAN Yu, WANG Jing, SUN Quan-sen, XIA De-shen   

  1. School of Computer Science and Technology, Nanjing University of Science and Technology,Nanjing 210094, China
  • Received:2010-06-19 Revised:2011-01-22 Online:2011-03-23 Published:2011-03-23
  • About author:潘瑜,博士生,研究方向:模式识别、图像处理、图像融合,E-mail: panyu_py@yahoo.com.cn;王静,博士生,研究方 向:航天遥感、图像处理、模式识别,E-mail:wangjing1195@126.com;孙权森,博士,教授,博导,研究方向:模式识别、图像处理、计算机视觉和机器学习,E-mail: qssun@126.com;夏德深,博士,教授,博导,研究方向:遥感图像处理、模式识别,E-mail:deshen_x@263.net
  • Supported by:

    国家自然科学基金(No.60773172);江苏省自然科学基金(No.BK2008411);教育部博士学科点基金(No.200802880017);江苏省高校自然科学研究项目基金(10KJB520011)资助

Abstract:

The fusion strategy of multi-scale image decomposition is the key point to improve the image fusion. To make full use of the information in the progress of multi-scale image decomposition, an image fusion strategy combined with image quality evaluation is proposed. This strategy includes two parts: first, the salient feature in the last level instead of current level is used to construct the weight which is necessary for the fusion of current level’s approximation images; second, the parameters of the image quality measurement such as spatial frequency and contrast are introduced for the fusion of detail images. The experiment demonstrates that, compared with current common fusion rules, the proposed fusion strategy has improved the quality of the fused image in subjective aspect as well in objective aspect.

Key words: image fusion, multi-scale decomposition, fusion strategy, approximation image, detail image

CLC Number: