Infrared and Visible Image Fusion Method Based on a Principal Component Analysis Network and Image Pyramid

The aim of infrared (IR) and visible image fusion is to generate a more informative for human observation or some other computer vision tasks. activity-level measurement weight assignment are two key parts in fusion. In this paper, we propose novel IR method based on the principal component analysis network (PCANet) an pyramid. Firstly, use lightweight deep learning network, PCANet, obtain images. obtained by PCANet has stronger representation ability focusing target perception detail description. Secondly, weights source images decomposed into multiple scales pyramid, weighted-average rule applied at each scale. Finally, fused reconstruction. effectiveness proposed algorithm was verified datasets with than eighty pairs test total. Compared nineteen representative methods, experimental results demonstrate that can achieve state-of-the-art both visual quality objective evaluation metrics.