基于融合技术的单幅红外图像增强方法

为更有效的提升红外图像的整体视觉效果并凸显其中的场景细节，提出了一个基于融合技术的单幅红外图像增强方法。首先, 利用现有图像增强方法构建多个富含不同有效分量的融合源。然后，以构建融合权重图的方式对融合源中的分量进行逐像素识别。最后，基于图像拉普拉斯金字塔模型，以分层融合的形式合成出增强后的图像。主观及客观对比实验结果证明了本方法具有良好的鲁棒性，及在视觉效果增强、有效信息增益方面的优势。

A Fusion-based Single Infrared Image Enhancement Method

In this paper, a fusion-based infrared image enhancement method is proposed aiming at improving the global visibility while increasing the scene details. The first procedure is to derive multiple fusion sources from the original input by applying different existing image enhancement techniques. Then, by constructing the fusion weight maps, important positive features within each fusion sources can be effectively recognized via a per-pixel strategy. Next, with the purpose of minimizing artifacts introduced by the fusion procedure, the Laplacian pyramid model are established for each fusion sources as well as the corresponding fusion weight maps. Finally, the enhanced image can be obtained via a per-layer fusion method. The comparison experimental results verify that the proposed method can produce results comparative to and even better than several existing state-of-the-art techniques with respect of the robustness and effectiveness.