A Multi-Layer Fusion Method for Infrared and Visible Images

The fusion of infrared and visible images serves as a pivotal technology to overcome the limited information-capturing capability of sensors in extreme environment. To address the various practical challenges, choosing an appropriate fusion strategy can be a prerequisite to ensure the performance of a fusion model. In this paper, a novel infrared and visible image fusion strategy is developed based on a multi-layer fusion framework. First, the image is decomposed into multiple parts based on the low-rank decomposition theory. Second, a fusion strategy tailored to the unique feature of each modality is employed. Then a multi-layer feature fusion strategy is proposed to enhance the fusion performance. Finally, the effectiveness of the proposed strategy is demonstrated in an experiment based on the TNO dataset, where the desired fusion result is achieved with marked enhancement in the visual quality.

Paper

Full text

PDF

A Multi-Layer Fusion Method for Infrared and Visible Images

Semantic Scholar · Engineering · 2023

Abstract

The fusion of infrared and visible images serves as a pivotal technology to overcome the limited information-capturing capability of sensors in extreme environment. To address the various practical challenges, choosing an appropriate fusion strategy can be a prerequisite to ensure the performance of a fusion model. In this paper, a novel infrared and visible image fusion strategy is developed based on a multi-layer fusion framework. First, the image is decomposed into multiple parts based on the low-rank decomposition theory. Second, a fusion strategy tailored to the unique feature of each modality is employed. Then a multi-layer feature fusion strategy is proposed to enhance the fusion performance. Finally, the effectiveness of the proposed strategy is demonstrated in an experiment based on the TNO dataset, where the desired fusion result is achieved with marked enhancement in the visual quality.

Similar papers

© 2026 NYSGPT2525 LLC