基于FRC框架的红外与可见光图像融合方法
作者:
作者单位:

1.上海电力大学;2.上海交通大学;3.上海极清慧视科技有限公司

作者简介:

通讯作者:

中图分类号:

TP391

基金项目:

国家重点基础研究发展计划(973计划),国家自然科学基金项目(面上项目,重点项目,重大项目),上海浦江人才计划,


Infrared and visible image fusion based on FRC algorithm
Author:
Affiliation:

1.Shanghai University of Electric Power;2.Shanghai UHDVISION Technology Co Ltd

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    摘 要:像素级图像融合方案多采用全局统一的融合规则,融合过程易导致区域层级信息丢失。对此,文中提出一种基于模糊区域特征(Fuzzy Regional Characteristics,FRC)框架的红外与可见光图像融合方法。首先,文中制定了一种模糊区域规则,该规则依据区域特征的模糊特性,对源图像依次施行区域分割、模糊判定和解模糊,全局图像的融合转化为三种不同属性区域的融合。然后,分别对目标区域、过渡区域和背景区域采用期望值最大(Expectation Maximization,EM)、加权平均融合、主成分分析(Principal Component Analysis,PCA)的融合策略进行融合。最后,融合后的区域图重构得到最终的融合图像。为验证FRC框架保护区域特征的有效性,文中还提出了一种区域特征一致性的度量指标(Regional Coherence,RCABF),经主观视觉与客观指标检验,融合图像在保持红外目标显著性的同时,具有清晰的背景细节,优于现有的红外与可见光图像融合方法,宜应用于红外目标的监控与分析。

    Abstract:

    Abstract:In pixel level image fusion schemes, global and unified fusion rules are often used, which results in the loss of regional level information. In this paper, a fusion method of infrared and visible images based on the frame of fuzzy regional characteristics (FRC) is proposed. First of all, a fuzzy region rule is established, according to the fuzzy characteristics of region features, the rule performs region segmentation, fuzzy judgment and defuzziness on the source image in turn. The fusion of global image is transformed into the fusion of three different attribute regions. Then, the target region, transition region and background region are fused by the fusion strategies of expectation maximization (EM), weighted average fusion and principal component analysis (PCA). Finally, the fused region image is reconstructed to get the final fused image. In order to verify the validity of FRC framework protection region feature, this paper also proposes a region feature consistency metric (RCABF), tested by subjective vision and objective indicators, the fusion image has clear background details while maintaining the salience of infrared target, which is superior to the existing fusion methods of infrared and visible image, and is suitable for infrared target monitoring and analysis.

    参考文献
    相似文献
    引证文献
引用本文
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2020-05-31
  • 最后修改日期:2020-10-08
  • 录用日期:2020-10-09
  • 在线发布日期:
  • 出版日期: