An image fusion algorithm based on image clustering theory: An image fusion algorithm based on image clustering theory: L. Zhao et al.

The fusion of infrared and visible light images aims to combine the advantages of both, in order to obtain a fused image with clear and textured targets. This article proposes a multi-scale image fusion method based on robust self-sparse fuzzy clustering. This method combines regularization techniques under Gaussian metrics to obtain sparse fuzzy membership degrees and automatically recognizes and eliminates noise interference. At the same time, connected component filtering based on a regional density balance strategy adaptively eliminates small domain interference, effectively extracting texture and target information from images. In terms of fusion strategy, we adopt morphological heterogeneity processing to balance the contrast between infrared and visible light, enriching the underlying image by utilizing the differential information between infrared and visible light images. Subsequently, the high-frequency components in the infrared and visible light images are linearly combined to maximize the preservation of infrared target information and visible light detail textures, achieving clear and rich fused images. Finally, through qualitative and quantitative analysis using existing advanced fusion methods on public datasets, it is demonstrated that our method not only successfully preserves target information in infrared images but also captures details and texture information in visible light images during the fusion process, resulting in clearer fusion images. The codes in this article are accessible at https://github.com/baidifeizi/FS.git. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.