Performance Comparison of Infrared and Visible Image Fusion Approaches

摘要

In this paper, we perform the performance comparison of image fusion methods from infrared and visible images. Here, we first use intensity-based registration method that image is aligned based on the relative intensity patterns for image registration. We also compare the performances of Sparse Representation (SR)-based and Multiscale Transform (MST)-based image fusion algorithms from OSU color-thermal database. Finally, we evaluate performance index by mutual information and entropy from final fusion images. It was found from the results that the presented image registration and fusion methods showed good performances.