[1]王 程,王 巍,杨 馨,等.基于区域特征优化及边缘增强的多聚焦图像融合[J].计算机技术与发展,2024,34(04):62-69.[doi:10. 3969 / j. issn. 1673-629X. 2024. 04. 010]
　WANG Cheng,WANG Wei,YANG Xin,et al.Multi-focus Image Fusion Based on Feature Optimization of Region and Edge Enhancement[J].,2024,34(04):62-69.[doi:10. 3969 / j. issn. 1673-629X. 2024. 04. 010]

点击复制

基于区域特征优化及边缘增强的多聚焦图像融合()

分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

34

期数:

2024年04期

页码:

62-69

栏目:

媒体计算

出版日期:

2024-04-10

文章信息/Info

Title:

Multi-focus Image Fusion Based on Feature Optimization of Region and Edge Enhancement

文章编号:

1673-629X(2024)04-0062-08

作者:

王 程¹ ; 王 巍² ; 杨 馨¹ ; 刘晓文¹

1. 中国人民公安大学 信息网络安全学院,北京 100038;

2. 中国人民公安大学 公安遥感应用工程技术研究中心,北京 100038

Author(s):

WANG Cheng1 ; WANG Wei2 ; YANG Xin1 ; LIU Xiao-wen1

1. School of Information and Cyber Security,People’ s Public Security University of China,Beijing 100038,China;

2. Remote Sensing Center of Public Security,People’s Public Security University of China,Beijing 100038,China

关键词:

多聚焦图像; 图像融合; 卷积神经网络; 特征优化; 边缘增强

Keywords:

multi-focus image; image fusion; convolutional neural network; feature optimization; edge enhancement

分类号:

TP301. 6

DOI:

10. 3969 / j. issn. 1673-629X. 2024. 04. 010

摘要:

针对多聚焦图像融合中存在的聚焦信息缺失、边缘特征表达不充分的问题,提出了一种基于聚焦区域特征优化及边缘增强的多聚焦图像融合方法。 首先,提出了一种四流向的聚焦特征增强网络,在图像编码过程中实现聚焦特征的增强与补全;其次,提出了一种新的基于空间频率(NSF) 的聚焦区域提取策略,实现对聚焦区域特征的高效提取;最后,提出了一种边缘特征增强模块,通过计算图像边缘梯度,实现对目标区域的边缘信息增强。 在 Lytro 数据集与 MFFW 数据集中与 8 种融合方法进行定性与定量对比。 实验结果表明,该融合方法较好地解决了多聚焦融合图像中聚焦特征表达不充分、纹理细节不完整和边缘模糊等问题,视觉效果显著。

Abstract:

In order to solve the problems of missing focus information and insufficient expression of edge details in multi-focus fusionimages,a multi-focus image fusion method based

on feature optimization of focus region and edge enhancement is proposed. Firstly,afour direction focusing feature enhancement network is proposed, which can enhance

and complete the focusing feature during imagecoding. Secondly,a new focus region extraction strategy based on spatial frequency ( NSF) is proposed,which can efficiently extract thefeatures of focus region in the coding process. Finally,an edge feature enhancement module is proposed,which can effectively calculatethe edge gradient of the

image to enhance the edge information of the target area. Qualitative and quantitative comparisons wereconducted with eight fusion methods in the Lytro dataset and the

MFFW dataset. The experimental results show the proposed fusionmethod can solve the problems of insufficient expression of focus features, incomplete texture details and fuzzy edges in multi - focusfusion images effectively,with remarkable visual effect.

常用功能