Abstract:

In the field of image fusion, it is very important to evaluate the quality of image fusion scientifically and specifically. This paper analyzes and summarizes the existing methods for objectively evaluating the quality of multiple image fusion, and conducts a targeted classification analysis based on the different points of reflection. The existing objective evaluation indexes are divided into three categories: the first is based on the statistical characteristics of the fusion image, the second is based on the objective evaluation index between the fusion image and the ideal reference image, the third is based on objective evaluation index between source images and the fusion image. This paper summarizes the existing objective evaluation methods and analyzes the latest objective methods of image fusion quality assessment. Finally, this paper analyzes and prospects the existing defects of the objective evaluation methods for image fusion quality and the future research trends, to further provide theoretical support for image fusion quality evaluation.

Key words: image fusion, subjective evaluation method, objective evaluation method, structural similarity

摘要：

在图像融合领域中，如何科学、有针对性地对图像融合质量进行有效评价具有十分重要的意义。分析和总结了目前已有的多种图像融合质量的客观评价方法，根据其反映侧重点的不同对其进行了针对性的分类分析。将目前已有的客观评价指标分为三大类：第一类是基于融合图像统计特征的客观评价指标；第二类是基于融合图像与理想参考图像的客观评价指标；第三类是基于源图像与融合图像的客观评价指标。分类归纳总结了目前已有的客观评价方法，对目前最新的图像融合质量客观评价方法进行了分析研究，并对目前已有的图像融合质量客观评价方法存在的缺陷和未来的研究趋势进行了分析与展望，进一步为图像融合质量评价提供理论支撑。

关键词: 图像融合, 主观评价方法, 客观评价方法, 结构相似性