Exploiting Digital Surface Models for Inferring Super-Resolution for Remotely Sensed Images
Not on the map. Why?
Abstract (arXiv)
Despite the plethora of successful Super-Resolution Reconstruction (SRR) models applied to natural images, their application to remote sensing imagery tends to produce poor results. Remote sensing imagery is often more complicated than natural images and has its peculiarities such as being of lower resolution, it contains noise, and often depicting large textured surfaces. As a result, applying non-specialized SRR models on remote sensing imagery results in artifacts and poor reconstructions. To address these problems, this paper proposes an architecture inspired by previous research work, introducing a novel approach for forcing an SRR model to output realistic remote sensing images: instead of relying on feature-space similarities as a perceptual loss, the model considers pixel-level information inferred from the normalized Digital Surface Model (nDSM) of the image. This strategy allows the application of better-informed updates during the training of the model which sources from a task (elevation map inference) that is closely related to remote sensing. Nonetheless, the nDSM auxiliary information is not required during production and thus the model infers a super-resolution image without any additional data besides its low-resolution pairs. We assess our model on two remotely sensed datasets of different spatial resolutions that also contain the DSM pairs of the images: the DFC2018 dataset and the dataset containing the national Lidar fly-by of Luxembourg. Based on visual inspection, the inferred super-resolution images exhibit particularly superior quality. In particular, the results for the high-resolution DFC2018 dataset are realistic and almost indistinguishable from the ground truth images.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 23 of its 55 references that are in the map. First posted on arXiv 9 May 2022.
Builds on
23 of 55 references.
References · 23
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
- 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
- 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
- 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
- 2016Densely Connected Convolutional Networkscited by 6,691
- 2022Generative Adversarial Networkscited by 5,603
- 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,590
- 2016Perceptual Losses for Real-Time Style Transfer and Super-Resolutioncited by 2,678
- 2018YOLOv3: An Incremental Improvementcited by 2,652
- 2016Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Networkcited by 1,898
- 2018Image Super-Resolution Using Very Deep Residual Channel Attention Networkscited by 1,256
11 more
- 2018ESRGAN: Enhanced Super-Resolution Generative Adversarial Networkscited by 1,250
- 2015Accurate Image Super-Resolution Using Very Deep Convolutional Networkscited by 902
- 2017Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolutioncited by 452
- 2014Understanding Deep Image Representations by Inverting Themcited by 399
- 2018The relativistic discriminator: a key element missing from standard GANcited by 234
- 2016Understanding How Image Quality Affects Deep Neural Networkscited by 118
- 2019How do neural networks see depth in single images?cited by 82
- 2019An Alternative Probabilistic Interpretation of the Huber Losscited by 43
- 2020Deep Learning for Change Detection in Remote Sensing Images: Comprehensive Review and Meta-Analysiscited by 34
- 2021A Review on Deep Learning in UAV Remote Sensingcited by 32
- 2020ESRGAN+ : Further Improving Enhanced Super-Resolution Generative Adversarial Networkcited by 27
Built on by
Citing papers per year: 2025: 1.
1 paper.