Liu Yang, Wang Wei, Li Weihe
School of Computer Science and Technology, Wuhan University of Science and Technology, Wuhan, China.
Hubei Province Key Laboratory of Intelligent Information Processing and Real-Time Industrial System, Wuhan University of Science and Technology, Wuhan, China.
PLoS One. 2025 Jun 3;20(6):e0324236. doi: 10.1371/journal.pone.0324236. eCollection 2025.
Pansharpening aims to combine the spatial information from high-resolution panchromatic (PAN) images with the spectral information from low-resolution multispectral (LRMS) images generating high-resolution multispectral (HRMS) images. While Convolutional Neural Networks (CNNs) have shown impressive performance in pansharpening tasks, their tendency to focus more on low-frequency information will lead to suboptimal preservation of high-frequency details, which are crucial for producing HRMS images. Recent studies have highlighted the significance of frequency domain information in pansharpening, but existing methods often consider the network as a whole, overlooking the unique abilities of different layers in capturing high-frequency components. This oversight can result in the loss of fine details and limit the overall performance of pansharpening. To overcome these limitations, we propose FIAN, a novel frequency information-adaptive network designed specifically for spatial-frequency domain pansharpening. FIAN introduces an innovative frequency information-adaptive filter module that can dynamically extract frequency-domain information at various frequencies, enabling the network to better capture and preserve high-frequency details during the pansharpening process. Furthermore, we have developed a frequency feature selection strategy to accurately extract the most relevant frequency-domain information, enhancing the network's representational power. Lastly, we present a multi-frequency information fusion module that effectively combines the frequency-domain information extracted by the filter at different frequencies with the spatial-domain information. We conducted extensive experiments on multiple benchmark datasets to evaluate the effectiveness of the proposed method. The experimental results demonstrate that our approach achieves competitive performance compared to state-of-the-art pansharpening methods.
全色锐化旨在将高分辨率全色（PAN）图像的空间信息与低分辨率多光谱（LRMS）图像的光谱信息相结合，以生成高分辨率多光谱（HRMS）图像。虽然卷积神经网络（CNN）在全色锐化任务中表现出了令人印象深刻的性能，但其更关注低频信息的倾向会导致高频细节的保存不理想，而高频细节对于生成HRMS图像至关重要。最近的研究强调了频域信息在全色锐化中的重要性，但现有方法通常将网络视为一个整体，忽略了不同层在捕获高频分量方面的独特能力。这种疏忽可能会导致精细细节的丢失，并限制全色锐化的整体性能。为了克服这些限制，我们提出了FIAN，一种专门为空间频域全色锐化设计的新型频率信息自适应网络。FIAN引入了一个创新的频率信息自适应滤波器模块，该模块可以动态提取不同频率的频域信息，使网络在全色锐化过程中能够更好地捕获和保留高频细节。此外，我们还开发了一种频率特征选择策略，以准确提取最相关的频域信息，增强网络的表征能力。最后，我们提出了一个多频率信息融合模块，该模块有效地将滤波器在不同频率提取的频域信息与空间域信息相结合。我们在多个基准数据集上进行了广泛的实验，以评估所提出方法的有效性。实验结果表明，与现有的全色锐化方法相比，我们的方法具有竞争力。
