Panchromatic Research Articles

Interpretable Model-Driven Deep Network for Hyperspectral, Multispectral, and Panchromatic Image Fusion.

Simultaneously fusing hyperspectral (HS), multispectral (MS), and panchromatic (PAN) images brings a new paradigm to generate a high-resolution HS (HRHS) image. In this study, we propose an interpretable model-driven deep network for HS, MS, and PAN image fusion, called HMPNet. We first propose a new fusion model that utilizes a deep before describing the complicated relationship between the HRHS and PAN images owing to their large resolution difference. Consequently, the difficulty of traditional model-based approaches in designing suitable hand-crafted priors can be alleviated because this deep prior is learned from data. We further solve the optimization problem of this fusion model based on the proximal gradient descent (PGD) algorithm, achieved by a series of iterative steps. By unrolling these iterative steps into several network modules, we finally obtain the HMPNet. Therefore, all parameters besides the deep prior are learned in the deep network, simplifying the selection of optimal parameters in the fusion and achieving a favorable equilibrium between the spatial and spectral qualities. Meanwhile, all modules contained in the HMPNet have explainable physical meanings, which can improve its generalization capability. In the experiment, we exhibit the advantages of the HMPNet over other state-of-the-art methods from the aspects of visual comparison and quantitative analysis, where a series of simulated as well as real datasets are utilized for validation.

A novel pansharpening model based on two parallel network architectures

ABSTRACT Pansharpening is an important technology for obtaining high-resolution multispectral (HRMS) images by fusing low-resolution multispectral (LRMS) images and high-resolution panchromatic (PAN) images. Although many pansharpening models have emerged by taking advantage of deep learning (DL) technology, there remains a pressing need to further assess pansharpening accuracy and stability when LRMS images with complex land-cover types. What’s more, these models often overlook the exploitation of PAN images’ inherent high-frequency information. To address these issues, we propose a pansharpening model combining multi-level and multi-scale network architectures. The multi-level network architecture is used to build spatial-spectral dependence on LRMS-PAN pairs, and strengthen the network’s feature capture capability by keeping the multi-level texture details. The multi-scale architecture is subsequently used to extract the spatial structure and deep texture of the PAN images at different scales. Downsampled experiments and real experiments in four standard datasets show that the proposed model achieves a state-of-the-art performance.

Kernel-guided injection deep network for blind fusion of multispectral and panchromatic images

Deep learning (DL)-based pansharpening methods show advantages in generating a high-resolution multispectral image (MS) by fusing low-resolution MS and high-resolution panchromatic (PAN) images. However, most current research only considers a single fixed kernel in the training process, thus they may fail to handle unknown degradations by using a single model for testing, i.e., blind pansharpening. Consequently, poor performance will occur when encountering the mismatched degradation problem. Therefore, building a single network that can tackle blind pansharpening is more realistic and challenging. To this end, we design an end-to-end blind pansharpening network composed of a kernel estimation sub-network and a fusion sub-network, named BPNet. In particular, different from existing DL-based methods, we control injected details by utilizing prior degradation information produced by kernel estimation sub-network, thus it can adapt to different types of degradation. Besides, we design an effective fusion sub-network with a strong interpretability, which injects details of PAN image into MS image by a learnable injection matrix. Especially, the estimated kernel that indicates the blur level of MS image is used to update the learnable injection matrix. To preserve important features well in the deep network, channel-Transformer is further utilized as a skip-connection in the fusion sub-network. With the combination of kernel estimation and fusion sub-networks, the proposed BPNet has great advantages in complex degradation situations, as the network’s adaptability to diverse unknown degradations is enhanced. Extensive experiments on simulated and real datasets demonstrate that BPNet outperforms other state-of-the-art methods in terms of visual results and objective quality analysis.

Pansharpening Based on Multimodal Texture Correction and Adaptive Edge Detail Fusion

Pansharpening refers to the process of fusing multispectral (MS) images with panchromatic (PAN) images to obtain high-resolution multispectral (HRMS) images. However, due to the low correlation and similarity between MS and PAN images, as well as inaccuracies in spatial information injection, HRMS images often suffer from significant spectral and spatial distortions. To address these issues, a pansharpening method based on multimodal texture correction and adaptive edge detail fusion is proposed in this paper. To obtain a texture-corrected (TC) image that is highly correlated and similar to the MS image, the target-adaptive CNN-based pansharpening (A-PNN) method is introduced. By constructing a multimodal texture correction model, intensity, gradient, and A-PNN-based deep plug-and-play correction constraints are established between the TC and source images. Additionally, an adaptive degradation filter algorithm is proposed to ensure the accuracy of these constraints. Since the TC image obtained can effectively replace the PAN image and considering that the MS image contains valuable spatial information, an adaptive edge detail fusion algorithm is also proposed. This algorithm adaptively extracts detailed information from the TC and MS images to apply edge protection. Given the limited spatial information in the MS image, its spatial information is proportionally enhanced before the adaptive fusion. The fused spatial information is then injected into the upsampled multispectral (UPMS) image to produce the final HRMS image. Extensive experimental results demonstrated that compared with other methods, the proposed algorithm achieved superior results in terms of both subjective visual effects and objective evaluation metrics.

An Adaptive Migration Collaborative Network for Multimodal Image Classification.

The multispectral (MS) and the panchromatic (PAN) images belong to different modalities with specific advantageous properties. Therefore, there is a large representation gap between them. Moreover, the features extracted independently by the two branches belong to different feature spaces, which is not conducive to the subsequent collaborative classification. At the same time, different layers also have different representation capabilities for objects with large size differences. In order to dynamically and adaptively transfer the dominant attributes, reduce the gap between them, find the best shared layer representation, and fuse the features of different representation capabilities, this article proposes an adaptive migration collaborative network (AMC-Net) for multimodal remote-sensing (RS) images classification. First, for the input of the network, we combine principal component analysis (PCA) and nonsubsampled contourlet transformation (NSCT) to migrate the advantageous attributes of the PAN and the MS images to each other. This not only improves the quality of images themselves, but also increases the similarity between the two images, thereby reducing the representational gap between them and the pressure on the subsequent classification network. Second, for the interaction on the feature migrate branch, we design a feature progressive migration fusion unit (FPMF-Unit) based on the adaptive cross-stitch unit of correlation coefficient analysis (CCA), which can make the network automatically learn the features that need to be shared and migrated, aiming to find the best shared-layer representation for multifeature learning. And we design an adaptive layer fusion mechanism module (ALFM-Module), which can adaptively fuse features of different layers, aiming to clearly model the dependencies among multiple layers for different sized objects. Finally, for the output of the network, we add the calculation of the correlation coefficient to the loss function, which can make the network converge to the global optimum as much as possible. The experimental results indicate that AMC-Net can achieve competitive performance. And the code for the network framework is available at: https://github.com/ru-willow/A-AFM-ResNet.

Multi-Frequency Spectral–Spatial Interactive Enhancement Fusion Network for Pan-Sharpening

The objective of pan-sharpening is to effectively fuse high-resolution panchromatic (PAN) images with limited spectral information and low-resolution multispectral (LR-MS) images, thereby generating a fused image with a high spatial resolution and rich spectral information. However, current fusion techniques face significant challenges, including insufficient edge detail, spectral distortion, increased noise, and limited robustness. To address these challenges, we propose a multi-frequency spectral–spatial interaction enhancement network (MFSINet) that comprises the spectral–spatial interactive fusion (SSIF) and multi-frequency feature enhancement (MFFE) subnetworks. The SSIF enhances both spatial and spectral fusion features by optimizing the characteristics of each spectral band through band-aware processing. The MFFE employs a variant of wavelet transform to perform multiresolution analyses on remote sensing scenes, enhancing the spatial resolution, spectral fidelity, and the texture and structural features of the fused images by optimizing directional and spatial properties. Moreover, qualitative analysis and quantitative comparative experiments using the IKONOS and WorldView-2 datasets indicate that this method significantly improves the fidelity and accuracy of the fused images.

Bayesian decision based fusion algorithm for remote sensing images

Remote sensing image fusion is dedicated to obtain a high-resolution multispectral (HRMS) image without spatial or spectral distortion compared to the single source image. In this paper, a novel fusion algorithm based on Bayesian estimation for remote sensing images is proposed from the new perspective of risk decisions. In this study, an observation model based on Bayesian estimation for remote sensing image fusion is constructed. Three categories of probabilities including prior, conditional and posterior probabilities are calculated after an intensity-hue-saturation (IHS) transformation is applied to the original low-resolution MS image. To obtain the desired HRMS image, with the corrected posterior probability, a fusion rule based on Bayesian decisions is designed to estimate which pixels to select from the panchromatic (PAN) image and the intensity component of the MS image. The selected pixels constitute a new component that will participate in an IHS inverse transformation to yield the fused image. Extensive experiments were performed on the Pleiades, WorldView-3, and IKONOS datasets, and the results demonstrate the effectiveness of the proposed method.

Selection of suitable fusional band combination from Sentinel-2A and UAV imagery for agricultural applications

ABSTRACT This study explores the fusion of Sentinel-2A (S2) and different bands of UAV imagery (Blue, Green, Red and Panchromatic (PAN)) to enhance spatial and spectral information for crop health monitoring. It assesses the quality of fused images, accuracy of the classified fused images and statistical analysis between the fused and S2 vegetation indices. Various fusion techniques are examined to select an optimal combination of bands. The quality of the fused images is evaluated using image quality assessment metrics and classification employs the Random Forest algorithm, assessing accuracy with metrics like F-score and Kappa Coefficient. Statistical analysis involves comparing vegetation indices from fused and S2 imagery. Notably, the fusion of UAV Green and Red bands with S2 imagery, using BT and PCA techniques, emerges as an effective combination for plant-level agricultural health monitoring. This research contributes to advancing precision agriculture techniques by leveraging multispectral imaging fusion for enhanced crop monitoring and management.

TemPanSharpening: A multi-temporal Pansharpening solution based on deep learning and edge extraction

The tradeoff among spatial, temporal, and spectral resolution of remote sensing (RS) images due to sensor properties limits the development of RS applications. Most image enhancement studies tend to focus on either spatio-temporal fusion or spatio-spectral fusion. As a more comprehensive solution, spatial–temporal-spectral fusion (STSF) is complicated but its potential is worth to be further explored. In this study, we propose a novel STSF method from the perspective of multi-temporal Pansharpening. Canny edge extraction is applied to Panchromatic (PAN) images to identify edges while avoiding the disruption of multi-temporal land cover changes. We then build a TemPanSharpening net (TPSnet) which only uses one high-spatio-low-spectra-temporal PAN and one low-spatio-high-spectra-temporal multispectral image as input. TPSnet follows a super-resolution structure and embeds two basic modules: residual-in-residual dense blocks (RRDB) and convolutional block attention module (CBAM). A series of interior ablation experiments were conducted on TPSnet and we also compared it with some representative spatio-temporal fusion, Pansharpening, and STSF algorithms. TPSnet presented satisfactory performance on complicated meter-level ground surfaces according to the quantitative evaluation result, and it demonstrated excellent robustness to land cover change.

Unsupervised Pan-Sharpening via Mutually Guided Detail Restoration

Pan-sharpening is a task that aims to super-resolve the low-resolution multispectral (LRMS) image with the guidance of a corresponding high-resolution panchromatic (PAN) image. The key challenge in pan-sharpening is to accurately modeling the relationship between the MS and PAN images. While supervised deep learning methods are commonly employed to address this task, the unavailability of ground-truth severely limits their effectiveness. In this paper, we propose a mutually guided detail restoration method for unsupervised pan-sharpening. Specifically, we treat pan-sharpening as a blind image deblurring task, in which the blur kernel can be estimated by a CNN. Constrained by the blur kernel, the pan-sharpened image retains spectral information consistent with the LRMS image. Once the pan-sharpened image is obtained, the PAN image is blurred using a pre-defined blur operator. The pan-sharpened image, in turn, is used to guide the detail restoration of the blurred PAN image. By leveraging the mutual guidance between MS and PAN images, the pan-sharpening network can implicitly learn the spatial relationship between the two modalities. Extensive experiments show that the proposed method significantly outperforms existing unsupervised pan-sharpening methods.

Progressive High-Frequency Reconstruction for Pan-Sharpening with Implicit Neural Representation

Pan-sharpening aims to leverage the high-frequency signal of the panchromatic (PAN) image to enhance the resolution of its corresponding multi-spectral (MS) image. However, deep neural networks (DNNs) tend to prioritize learning the low-frequency components during the training process, which limits the restoration of high-frequency edge details in MS images. To overcome this limitation, we treat pan-sharpening as a coarse-to-fine high-frequency restoration problem and propose a novel method for achieving high-quality restoration of edge information in MS images. Specifically, to effectively obtain fine-grained multi-scale contextual features, we design a Band-limited Multi-scale High-frequency Generator (BMHG) that generates high-frequency signals from the PAN image within different bandwidths. During training, higher-frequency signals are progressively injected into the MS image, and corresponding residual blocks are introduced into the network simultaneously. This design enables gradients to flow from later to earlier blocks smoothly, encouraging intermediate blocks to concentrate on missing details. Furthermore, to address the issue of pixel position misalignment arising from multi-scale features fusion, we propose a Spatial-spectral Implicit Image Function (SIIF) that employs implicit neural representation to effectively represent and fuse spatial and spectral features in the continuous domain. Extensive experiments on different datasets demonstrate that our method outperforms existing approaches in terms of quantitative and visual measurements for high-frequency detail recovery.

Nonreference object-based pansharpening quality assessment

Pansharpening involves the fusion of panchromatic (PAN) and multispectral (MS) images to obtain a high-resolution image with enhanced spectral and spatial information. Assessing the quality of the resulting fused image poses a challenge due to the absence of a high-resolution reference image. Numerous methods have been proposed to address this, from assessing quality at reduced resolution to full-resolution evaluations. Many existing approaches are pixel-based, where quality metrics are applied and averaged on individual pixels. In this article, we introduce a novel object-based method for assessing the quality of pansharpened images at full resolution. In object-based quality assessment methods, the reaction of different areas of the fused image to the fusion process is reflected. Our approach revolves around extracting objects from the given image and evaluating extracted objects. By doing so, the distinct responses of different objects within the fused image to the fusion process are captured. The proposed method leverages a unique object extraction technique known as segmentation by nearest neighbor (SNN) to extract objects of the MS image. This method extracts the objects based on the image’s characteristics without any requirement for parameter tuning. These extracted objects are then mapped onto both PAN and fused images. The proposed spectral index measures the spectral homogeneity of the fused image’s objects and the proposed spatial index measures the injected spatial content from the PAN image to the fused image’s objects. Experimental results underscore the robustness and reliability of the proposed method. Additionally, by visualizing distortion values on object-maps, we gain insights into fusion quality across diverse areas within the scene.

DUCD: Deep Unfolding Convolutional-Dictionary network for pansharpening remote sensing image

The goal of pansharpening methods is to complement the spectral and spatial information contained in Multi-spectral (MS) and panchromatic (PAN) images to obtain the desired High-resolution multispectral (HRMS) image. The existing majority of pansharpening methods either extract feature information separately from the MS image and PAN image, or extract feature information after concatenating the MS image and PAN image. However, the entire extraction process lacks the utilization of complementary information and tends to generate redundant information, thereby leading to the loss of certain important information during the extraction process, which in turn affects the overall performance. In order to better utilize the complementary information between the MS image and PAN image and enhance the interpretability of the network, we propose the Deep Unfolding Convolutional-Dictionary Network (DUCD) for pansharpening in this paper. This network fully integrates complementary information between the MS image and PAN image to generate the final fused image. The entire network structure consists of two parts: The encoder and the decoder. In the encoder part of the network, we clarify the common and unique feature information between MS and PAN images by constructing an observation model. Simultaneously, we use the approximate gradient algorithm to continuously optimize the model and iteratively unfold it into a deep network structure. In the decoder part of the network, we concatenate the obtained common and specific information from MS and PAN images and pass them through convolutional and activation layers. Subsequently, they are input into the introduced Frequency Domain-based Transformer (FDT) module and an information-lossless inversible neural network(INN). This provides a more efficient method for establishing long-range dependency relationships between feature extraction and feature fusion. To demonstrate the effectiveness of our proposed method, we conduct extensive experiments on three benchmark datasets QB, GF2 and WV3. Experimental results show that our method outperforms the current SOTA Pansharpening methods in terms of performance.

GSA-SiamNet: A Siamese Network with Gradient-Based Spatial Attention for Pan-Sharpening of Multi-Spectral Images

Pan-sharpening is a fusion process that combines a low-spatial resolution, multi-spectral image that has rich spectral characteristics with a high-spatial resolution panchromatic (PAN) image that lacks spectral characteristics. Most previous learning-based approaches rely on the scale-shift assumption, which may not be applicable in the full-resolution domain. To solve this issue, we regard pan-sharpening as a multi-task problem and propose a Siamese network with Gradient-based Spatial Attention (GSA-SiamNet). GSA-SiamNet consists of four modules: a two-stream feature extraction module, a feature fusion module, a gradient-based spatial attention (GSA) module, and a progressive up-sampling module. In the GSA module, we use Laplacian and Sobel operators to extract gradient information from PAN images. Spatial attention factors, learned from the gradient prior, are multiplied during the feature fusion, up-sampling, and reconstruction stages. These factors help to keep high-frequency information on the feature map as well as suppress redundant information. We also design a multi-resolution loss function that guides the training process under the constraints of both reduced- and full-resolution domains. The experimental results on WorldView-3 satellite images obtained in Moscow and San Juan demonstrate that our proposed GSA-SiamNet is superior to traditional and other deep learning-based methods.

A Spatio-Spectral Fusion Method for Hyperspectral Images Using Residual Hyper-Dense Network.

Spatio-spectral fusion of panchromatic (PAN) and hyperspectral (HS) images is of great importance in improving spatial resolution of images acquired by many commercial HS sensors. DenseNets have recently achieved great success for image super-resolution because they facilitate gradient flow by concatenating all the feature outputs in a feedforward manner. In this article, we propose a residual hyper-dense network (RHDN) that extends the DenseNet to solve the spatio-spectral fusion problem. The overall structure of the proposed RHDN method is a two-branch network, which allows the network to capture the features of HS images within and outside the visible range separately. At each branch of the network, a two-stream strategy of feature extraction is designed to process PAN and HS images individually. A convolutional neural network (CNN) with cascade residual hyper-dense blocks (RHDBs), which allows direct connections between the pairs of layers within the same stream and those across different streams, is proposed to learn more complex combinations between the HS and PAN images. The residual learning is adopted to make the network efficient. Extensive benchmark evaluations well demonstrate that the proposed RHDN fusion method yields significant improvements over many widely accepted state-of-the-art approaches.

Estimation of spectral similarities utilizing segmented regions' probability distribution in the block-optimized pan-sharpened image for material classification.

Pan-sharpening is an image fusion approach that combines the spectral information in multispectral (MS) images with the spatial properties of PAN (Panchromatic) images. This vital technique is used in categorization, detection, and other remote sensing applications. In the first step, the article focuses on increasing the finer spatial details in the MS image with PAN images using two levels of fusion without causing spectral deterioration. The suggested fusion method efficiently utilizes image transformation techniques and spatial domain image fusion methods. The luminance component of MS images typically contains spatial features that are not as detailed as the PAN images. A multiscale transform is applied to the intensity/luminance component and PAN image to introduce features into the intensity component. In the first level of processing, coefficients obtained from the non-subsampled contourlet transform are subjected to particle swarm optimization weighted block-based fusion. The second level of fusion is carried out using the concept of spatial frequency to reduce spectral distortion. Numerous reference and non-reference parameters are used to evaluate the sharpened image's quality. In the next step, the article focuses on designing an evaluation metric for analysing spectral distortion based on the Bhattacharyya coefficient and distance. The Bhattacharyya coefficient and distance are calculated for each segmented region to assess the sharpened images' quality. Spectral degradation analysis using proposed techniques can also be useful for analysing materials in the segmented regions. The research findings demonstrate that the spatial features of fused images obtained from the proposed technique increased with the least spectral degradation.

A Comprehensive Assessment of the Pansharpening of the Nighttime Light Imagery of the Glimmer Imager of the Sustainable Development Science Satellite 1

The Sustainable Development Science Satellite 1 (SDGSAT-1) satellite, launched in November 2021, is dedicated to providing data detailing the “traces of human activities” for the implementation of the United Union’s 2030 Agenda for Sustainable Development and global scientific research. The glimmer imager (GI) that is equipped on SDGSAT-1 can provide nighttime light (NL) data with a 10 m panchromatic (PAN) band and red, green, and blue (RGB) bands of 40 m resolution, which can be used for a wide range of applications, such as in urban expansion, population studies of cities, and economics of cities, as well as nighttime aerosol thickness monitoring. The 10 m PAN band can be fused with the 40 m RGB bands to obtain a 10 m RGB NL image, which can be used to identify the intensity and type of night lights and the spatial distribution of road networks and to improve the monitoring accuracy of sustainable development goal (SDG) indicators related to city developments. Existing remote sensing image fusion algorithms are mainly developed for daytime optical remote sensing images. Compared with daytime optical remote sensing images, NL images are characterized by a large amount of dark (low-value) pixels and high background noises. To investigate whether daytime optical image fusion algorithms are suitable for the fusion of GI NL images and which image fusion algorithms are the best choice for GI images, this study conducted a comprehensive evaluation of thirteen state-of-the-art pansharpening algorithms in terms of quantitative indicators and visual inspection using four GI NL datasets. The results showed that PanNet, GLP_HPM, GSA, and HR outperformed the other methods and provided stable performances among the four datasets. Specifically, PanNet offered UIQI values ranging from 0.907 to 0.952 for the four datasets, whereas GSA, HR, and GLP_HPM provided UIQI values ranging from 0.770 to 0.856. The three methods based on convolutional neural networks achieved more robust and better visual effects than the methods using multiresolution analysis at the original scale. According to the experimental results, PanNet shows great potential in the fusion of SDGSAT-1 GI imagery due to its robust performance and relatively short training time. The quality metrics generated at the degraded scale were highly consistent with visual inspection, but those used at the original scale were inconsistent with visual inspection.

Triple disentangled network with dual attention for remote sensing image fusion

Remote sensing applications, such as detection and recognition of objects, need a large number of high spatial resolution (HR) images. As a feasible technique, image fusion is considered to produce HR images. However, most of remote sensing image fusion methods based on deep neural network (DNN) are mainly limited by the extraction of spatial and spectral features. These methods ignore the redundancy among these features, which causes spatial and spectral distortions in the fused image. In this paper, we propose a novel image fusion method based on a triple disentangled network (TDNet) with dual attention to reduce the redundancy in the extracted features. In the proposed method, it is assumed that the information in panchromatic (PAN) and low spatial resolution multispectral (LRMS) images can be encoded as the spatial, spectral, and common features. Specifically, we construct a triple-stream network to extract these features. To efficiently model the spatial and spectral information in PAN and LRMS images, local–global attention and interdependency attention are designed and integrated into the network. Then, the redundancy among these features is reduced by disentangled learning, in which these features are recombined to reconstruct the PAN and LRMS images. Besides, these features should complement each other. So, we utilize the maximal coding rate reduction to balance the redundancy and complementarity among them. Finally, all features are recombined to synthesize the high spatial resolution multispectral image. The experimental results demonstrate that the proposed TDNet has a superior performance in terms of qualitative and quantitative evaluations. The code link is https://github.com/RSMagneto/TDNet.

CrossDiff: Exploring Self-SupervisedRepresentation of Pansharpening via Cross-Predictive Diffusion Model.

Fusion of a panchromatic (PAN) image and corresponding multispectral (MS) image is also known as pansharpening, which aims to combine abundant spatial details of PAN and spectral information of MS images. Due to the absence of high-resolution MS images, available deep-learning-based methods usually follow the paradigm of training at reduced resolution and testing at both reduced and full resolution. When taking original MS and PAN images as inputs, they always obtain sub-optimal results due to the scale variation. In this paper, we propose to explore the self-supervised representation for pansharpening by designing a cross-predictive diffusion model, named CrossDiff. It has two-stage training. In the first stage, we introduce a cross-predictive pretext task to pre-train the UNet structure based on conditional Denoising Diffusion Probabilistic Model (DDPM). While in the second stage, the encoders of the UNets are frozen to directly extract spatial and spectral features from PAN and MS images, and only the fusion head is trained to adapt for pansharpening task. Extensive experiments show the effectiveness and superiority of the proposed model compared with state-of-the-art supervised and unsupervised methods. Besides, the cross-sensor experiments also verify the generalization ability of proposed self-supervised representation learners for other satellite datasets. Code is available at https://github.com/codgodtao/CrossDiff.

Feature Pyramid Fusion Network for Hyperspectral Pansharpening.

Hyperspectral (HS) pansharpening aims at fusing an observed HS image with a panchromatic (PAN) image, to produce an image with the high spectral resolution of the former and the high spatial resolution of the latter. Most of the existing convolutional neural networks (CNNs)-based pansharpening methods reconstruct the desired high-resolution image from the encoded low-resolution (LR) representation. However, the encoded LR representation captures semantic information of the image and is inadequate in reconstructing fine details. How to effectively extract high-resolution and LR representations for high-resolution image reconstruction is the main objective of this article. In this article, we propose a feature pyramid fusion network (FPFNet) for pansharpening, which permits the network to extract multiresolution representations from PAN and HS images in two branches. The PAN branch starts from the high-resolution stream that maintains the spatial resolution of the PAN image and gradually adds LR streams in parallel. The structure of the HS branch remains highly consistent with that of the PAN branch, but starts with the LR stream and gradually adds high-resolution streams. The representations with corresponding resolutions of PAN and HS branches are fused and gradually upsampled in a coarse to fine manner to reconstruct the high-resolution HS image. Experimental results on three datasets demonstrate the significant superiority of the proposed FPFNet over the state-of-the-art methods in terms of both qualitative and quantitative comparisons.