全色图像锐化是遥感数据处理领域的一个基础性问题，在物体分类、目标识别等方面具有重要的研究意义和应用价值。 近年来, 深度学习在自然语言处理、计算机视觉等领域取得了巨大进展, 也推动了像素级全色图像锐化技术的发展。 本文提出从经典方式和协同方式两个方面对深度学习在全色图像锐化中的研究进行系统的综述，并在此基础上进行前景展望。首先，给出全色图像锐化常用的数据集和全色图像锐化的质量评价指标。接着，从经典方式与协同方式两个方面对基于深度学习的全色图像锐化最新研究成果进行分门别类的介绍，并进行算法性能的对比、分析和归纳。然后，对全色图像锐化的三个主要应用领域进行分析。最后，本文探讨了基于深度学习的全色图像锐化的五个未来研究方向。

Pansharpening is a fundamental problem in the field of remote sensing data processing. It has important research significance and application value in classification and object recognition. In recent years, deep learning has made great progress in natural language processing, computer vision and other fields, and promoted the development of pansharpening technology. This paper presents a systematic review on deep learning-based pixel-level pansharpening methods from two perspectives of classical mode and collaborative mode, and makes a prospect on this basis. Firstly, the widely used data sets and quality evaluation indexes of pansharpening are given. Then, from two aspects of classical and collaborative methods, the latest research results of pansharpening based on deep learning are introduced, analyzed and mined. Afterward, three main application fields of pansharpening are analyzed. Finally, this paper discusses five future research directions of deep learning-based pansharpening methods.