红外与可见光图像融合研究

Research on Infrared and Visible Image Fusion

红外与可见光图像融合是指将同一场景下的相机获取的图像与红外传感器获得的图像融合为一张图像。融合图像具备源图像的重要信息,尽量无冗余信息。融合图像在广泛应用在计算机视觉、农业、遥感、医学等领域。本文主要对其研究进展做整理,使感兴趣学者快速掌握红外与可见光图像融合研究脉络。首先将研究算法主要分为:多尺度变换、稀疏表示、神经网络以及其他方法等几大类。接着对每类进行详细探讨,然后对目前主流算法做一些优劣评价与总结。最后对本研究领域未来值得突破点进行展望。

Visible and infrared image fusion refers to the fusion of images obtained by cameras and infrared sensors in the same scene into one image. The fused image has important information from the source image, and there should be as little redundant information as possible. Fusion images are widely used in many fields such as medical, computer vision, agriculture, and remote sensing. This article mainly summarizes the research progress of visible and infrared image fusion, enabling interested scholars to quickly grasp the research context. Firstly, the algorithms are mainly divided into several categories: multi-scale transformation, sparse representation, neural networks, and other methods. Secondly, a detailed discussion will be conducted on each category, followed by some evaluation and summary of the current mainstream algorithms. Finally, prospects for future breakthroughs in this research field.