CLOUDTRAN: CLOUD REMOVAL FROM MULTITEMPORAL SATELLITE IMAGES USING AXIAL TRANSFORMER NETWORKS

Abstract. We present a method for cloud-removal from satellite images using axial transformer networks. The method considers a set of multitemporal images in a given region of interest together with the corresponding cloud masks, and delivers a cloud-free image for a specific day of the year. We propose the combination of an encoder-decoder model employing axial attention layers for the estimation of the low-resolution cloud-free image, together with a fully parallel upsampler that reconstructs the image at full resolution. The method is compared with various baselines and state-of-the-art methods on two Sentinel-2 datasets, showing significant improvements across multiple standard metrics used for image quality assessment.

• Publication year

  2022

• Venue

  The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences

• Publication date

  2022-05-30

• Fields of study

  Not labeled

• Identifiers
  DOI 10.5194/isprs-archives-xliii-b2-2022-1125-2022
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• SEN12MS-CR-TS: A Remote Sensing Data Set for Multi-modal Multi-temporal Cloud Removal

  2022influential reference
• Cloud Mask Intercomparison eXercise (CMIX): An evaluation of cloud masking algorithms for Landsat 8 and Sentinel-2

  2022cited by this paper
• Colorization Transformer

  2021cited by this paper
• FuseFormer: Fusing Fine-Grained Information in Transformers for Video Inpainting

  2021cited by this paper
• Sentinel-1 and Sentinel-2 Spatio-Temporal Data Fusion for Clouds Removal

  2021cited by this paper
• SACTNet: Spatial Attention Context Transformation Network for Cloud Removal

  2021cited by this paper
• Vision Transformers for Remote Sensing Image Classification

  2021cited by this paper
• A Deep Multitask Learning Framework Coupling Semantic Segmentation and Fully Convolutional LSTM Networks for Urban Change Detection

  2021cited by this paper
• Cloud Removal for Remote Sensing Imagery via Spatial Attention Generative Adversarial Network

  2020influential reference
• Axial Attention in Multidimensional Transformers

  2019cited by this paper
• Free-Form Video Inpainting With 3D Gated Convolution and Temporal PatchGAN

  2019cited by this paper
• A Review of Change Detection in Multitemporal Hyperspectral Images: Current Techniques, Applications, and Challenges

  2019cited by this paper
• Vision Based Modeling of Plants Phenotyping in Vertical Farming under Artificial Lighting

  2019cited by this paper
• Cloud-Gan: Cloud Removal for Sentinel-2 Imagery Using a Cyclic Consistent Generative Adversarial Networks

  2018cited by this paper
• Generating High Fidelity Images with Subscale Pixel Networks and Multidimensional Upscaling

  2018cited by this paper
• Detailed Land Cover Mapping from Multitemporal Landsat-8 Data of Different Cloud Cover

  2018cited by this paper
• Masked Autoregressive Flow for Density Estimation

  2017cited by this paper
• Attention is All you Need

  2017influential reference
• Cloud Removal From Multitemporal Satellite Images Using Information Cloning

  2013cited by this paper
• Image quality assessment: from error visibility to structural similarity

  2004cited by this paper

• Supervised Alignment Perception and Modality Difference Selection Enhanced SAR-Optical Fusion for Cloud Removal

  2025cites this paper
• A Spatial–Frequency Combined Transformer for Cloud Removal of Optical Remote Sensing Images

  2025cites this paper
• MSC-GAN: A Multistream Complementary Generative Adversarial Network With Grouping Learning for Multitemporal Cloud Removal

  2025cites this paper
• TCME: Thin Cloud Removal Network for Optical Remote Sensing Images Based on Multidimensional Features Enhancement

  2024cites this paper
• Cascaded Memory Network for Optical Remote Sensing Imagery Cloud Removal

  2024cites this paper
• Multiscale Restoration of Missing Data in Optical Time-Series Images With Masked Spatial–Temporal Attention Network

  2024cites this paper
• CloudTran++: Improved Cloud Removal from Multi-Temporal Satellite Images Using Axial Transformer Networks

  2024cites this paper
• Cloud-EGAN: Rethinking CycleGAN From a Feature Enhancement Perspective for Cloud Removal by Combining CNN and Transformer

  2023cites this paper
• Cloudformer: A Cloud-Removal Network Combining Self-Attention Mechanism and Convolution

  2022cites this paper
• PLFM: Pixel-Level Merging of Intermediate Feature Maps by Disentangling and Fusing Spatial and Temporal Data for Cloud Removal

  2022cites this paper