Infrared and Visible Image Fusion and Image Quality Evaluation

Abstract

Fusion of infrared and visible images can help extracting the characteristics of images from different band images while enhancing people’s recognition and understanding ability of the image in the field of detection. As a result, it has attracted widespread attention in the military and civil areas. The traditional fusion method of infrared and visible images cannot effectively highlight the infrared target information while retaining the texture details of visible image. To this end, a new fusion method is proposed to enhance the fusion effect of the infrared and visible images. Firstly, the source images are decomposed into texture detail layers, edge layers and a base layer through a mixed information decomposition method. The textural detail layers and edge layers are fused using a perceptual parameter selection method, the base layer is fused through a sparse representation based method. Finally, the final fused image can be reconstructed by the fused layers. In order to verify the effectiveness of the proposed method, the method is compared with the typical fusion methods. The experimental results of the former turn out to be better in terms of both subjective vision and objective evaluation indicators.