Amplifying the imaging power of digital sky surveys with space telescopes data and generative AI

Sai Teja Erukude; Lior Shamir

astro-ph.IMAug 21, 2026

Amplifying the imaging power of digital sky surveys with space telescopes data and generative AI

Authors: Sai Teja Erukude, Lior Shamir

Organizations: Kansas State University, Manhattan, KS, 66506, USA

Abstract

While Digital sky surveys provide excellent throughput of image data and can cover a large footprint, their imaging power is normally inferior to that of space-based telescopes. Space-based telescopes, on the other hand, provide excellent imaging power and can image the deep Universe, but cannot provide the same throughput as advanced ground-based sky surveys. Here, we utilize generative AI to elevate the quality of galaxy images taken by ground-based telescopes to the level of details enabled by space telescopes. The solution is based on the nature of galaxy shapes, allowing generative AI trained on space-based images to convert weak signal into detailed and clear galaxy images. The method allows for combining the high throughput of ground-based sky surveys with the image quality of space-based telescopes. The source code for the method is available, as well as paired training data and a catalog of 63,202 galaxy images enhanced by the proposed method. We also provide a software tool that encapsulates the entire pipeline and the custom generative AI model to generate galaxy images with enhanced quality.

Figures & tables

Figure 1: Pairs of training images used to train the cGAN. Each pair includes an image taken by the DESI Legacy Survey (left) and an image of the same object taken by the HST (right).

Figure 2: Overview of the data pre-processing pipeline, illustrating the inputs, the output, and key transformation steps involved.

Figure 3: U-Net architecture of the generator used in the enhancement of ground-based telescope images.

Figure 4: PatchGAN-based architecture of the discriminator.

Component	Setting
Model architecture	Pix2Pix conditional GAN (U-Net generator, PatchGAN discriminator) ( Isola et al., 2018 )
Generator backbone	U-Net encoder C64–C512; decoder CD512–C64 ( Ronneberger et al., 2015 )
Discriminator	PatchGAN, C64–C512, Sigmoid patch output ( Demir and Unal, 2018 )
Input resolution	$256\times 256\times 3$ RGB tiles (ground/space pairs)
Dataset type	Paired image translation (Ground $\rightarrow$ Space telescope)
Normalization	Pixel values scaled from [0, 255] to [-1, 1]

Table 1: Training configuration summary for the proposed Pix2Pix-based generative AI framework.

Component	Specification
Operating System	Debian GNU/Linux 12 (bookworm)
Kernel Version	6.1.0-28-amd64
CPU Model	2 × Intel Xeon Gold 5317, 3.0 GHz
CPU Cores / Threads	48 cores / 96 threads
System Memory	125 GiB RAM
Primary Storage	223 GB SSD (system)

Table 2: Hardware and software configuration of the experimental server.

Model	Trainable	Non-trainable	Total
	parameters	parameters	parameters
Generator	54,419,459	9,856	54,429,315
/ U-Net
Discriminator	6,965,441	2,816	6,968,257
/ PatchGAN
Full cGAN	54,422,275	6,975,297	61,397,572

Table 3: Number of parameters in the models.

\hat{x}_{B}\leftarrow G(x_{A})

Algorithm 1 Training procedure for the Pix2Pix conditional GAN model.

\hat{x}_{B}\leftarrow G^{\ast}(x_{A})

Algorithm 2 Inference Procedure for Enhancing Ground Telescope Images

Figure 5: Examples of ground-based telescope images, the space-based images of the same objects, and the images generated by the GAN from the ground-based images. The residual maps are created from the differences between generated images and the space telescope images. The objects are at coordinates $(\alpha=150.6532946^{o},\delta=1.6253657^{o})$ , $(\alpha=150.3196276^{o},\delta=1.7540747^{o})$ , $(\alpha=150.3742923^{o},\delta=1.6191228^{o})$ , $(\alpha=150.0445515^{o},\delta=1.6114927^{o})$ .

Figure 6: Examples from the catalog where the model demonstrates outstanding performance. Even when the input image lacks visible details of the galaxy, the model generates an enhanced version that closely matches the ground truth. The residual maps also show the good agreement between the generated image and space telescope image. The objects are at coordinates $(\alpha=149.68237^{o},\delta=2.4202963^{o})$ , $(\alpha=149.97054^{o},\delta=2.6837897^{o})$ , $(\alpha=149.8200861^{o},\delta=2.4452223^{o})$ , $(\alpha=149.8254184^{o},\delta=2.1643785^{o})$ .

Figure 7: Catalog examples highlighting cases where the model underperforms. These issues might be addressed by training the model on a larger and more diverse set of images. The objects are at coordinates $(\alpha=150.7538768^{o},\delta=1.957399^{o})$ , $(\alpha=150.7714556^{o},\delta=2.0616442^{o})$ , $(\alpha=150.7619236^{o},\delta=2.6261117^{o})$ , $(\alpha=150.7552252^{o},\delta=2.2725501^{o})$ .

Descriptors		$\overline{\text{Relative difference}}$
Major axis (A)		0.056
Minor axis (B)		0.052
Position angle ( $\theta$ )		0.031
Elongation (A/B)		0.05
Ellipticity 1-(B/A)		0.05
Broad morphology		0.06

Table 4: The differences between galaxy descriptors computed from the space telescope images and the same attributed computed from the AI-enhanced images of the same galaxies.

Figure 8: Sample enhanced images from the catalog of enhanced galaxy images.

Figure 9: Example images from the “galaxy_comparisons” of the catalog.

Figure 10: Modules and data flow involved in the galaxy enhancer tool.

Figure 11: Terminal output illustrating each stage of the galaxy enhancer execution.

Figure 12: Examples of enhanced SDSS galaxies in comparison to JWST deep field.

Explore similar work

FluxFlow: Conservative Flow-Matching for Astronomical Image Super-Resolution
May 5, 2026Shuhong Liu, Xining Ge, Ziteng Cui +8Flow Matching Image Super-Resolution
1The University of Tokyo · 2I2WM · 3Tohoku University +1
Accelerating Redshift-Conditioned Galaxy Image Synthesis with One-step Generative Modeling
May 17, 2026Tianyue Yang, Sandro Tacchella, Xiao XueConditional Image Generation One-Step Generative Modeling
The Center for Computational Science University College London · Cavendish Laboratory Kavli Institute for Cosmology University of Cambridge
Space Generative AI with Solar Energy Harvesting
Sep 1, 2026Jierui Zhang, Jianhao Huang, Zhanwei Wang +1Wireless Communications
Department of Electrical and Computer Engineering, The University of Hong Kong (HKU), Hong Kong SAR, China

May 5, 2026J/K move · Enter open · S save

FluxFlow: Conservative Flow-Matching for Astronomical Image Super-Resolution

Shuhong Liu, Xining Ge, Ziteng Cui, Liuzhuozheng Li+7

1The University of Tokyo · 2I2WM · 3Tohoku University+1

Flow Matching Image Super-Resolution

Open paper AskPDF arXiv

Ground-to-space astronomical super-resolution requires recovering space-quality images from ground-based observations that are simultaneously limited by pixel sampling resolution and atmospheric seeing, which imposes a stochastic, spatially varying PSF that cannot be resolved through upsampling alone. Existing methods rely on synthetic training pairs that fail to capture real atmospheric statistics and are prone to either over-smoothed reconstructions or hallucination sources with no physical counterpart in the observed sky. We propose FluxFlow, a conservative pixel-space flow-matching framework that incorporates observation uncertainty and source-region importance weights during training, and a training-free Wiener-regularized test-time correction to suppress hallucination sources while preserving recovered detail. We further construct the DESI--HST Dataset, the large-scale real-world benchmark comprising 19,500 real co-registered ground-to-space image pairs with real atmospheric PSF variation. Experiments demonstrate that FluxFlow consistently outperforms existing baseline methods in both photometric and scientific accuracy.

Abstract

Explore similar work

FluxFlow: Conservative Flow-Matching for Astronomical Image Super-Resolution

Accelerating Redshift-Conditioned Galaxy Image Synthesis with One-step Generative Modeling

Space Generative AI with Solar Energy Harvesting