S 3 R-GS: Saliency-Guided Gaussian Splatting for Arbitrary-Scale Spacecraft Image Super-Resolution.
Saved in:
| Title: | S 3 R-GS: Saliency-Guided Gaussian Splatting for Arbitrary-Scale Spacecraft Image Super-Resolution. |
|---|---|
| Authors: | Liu, Chuyang1 (AUTHOR), Wu, Liangyi2 (AUTHOR), Liu, Kai3 (AUTHOR), Chen, Luyang4 (AUTHOR), Wei, Xin1,4 (AUTHOR) weixin@xidian.edu.cn, Yang, Xi2,4 (AUTHOR) |
| Source: | Remote Sensing. May2026, Vol. 18 Issue 10, p1585. 23p. |
| Subjects: | Image reconstruction, Clustering algorithms, Pose estimation (Computer vision), Image representation, High resolution imaging |
| Abstract: | Highlights: What are the main findings? This study proposes S3R-GS, an arbitrary-scale super-resolution model that reconstructs spacecraft images at any magnification factor using a single network. The model uses spatial clustering to assign Gaussian kernels strictly to the spacecraft body, avoiding wasteful computations on the featureless deep-space background. What are the implications of the main finding? Relying on explicit 2D Gaussian shapes rather than point-wise regression allows S3R-GS to maintain sharp details even when extrapolated far beyond typical scales (up to ×12). Tests across different datasets demonstrate that stripping away background noise leads to much cleaner object boundaries, which directly benefits subsequent tasks like optical pose estimation. High-resolution images of non-cooperative spacecraft are essential for on-board autonomous operations. Hardware bandwidth limits and continuously changing observation distances mean that a practical super-resolution (SR) system must handle arbitrary, non-integer magnification factors without retraining, a setting known as arbitrary-scale SR (ASSR). Recent 2D Gaussian splatting (2DGS) methods represent image content with explicit anisotropic Gaussian primitives and render at any continuous coordinate, offering substantially faster inference than implicit neural representation (INR) approaches. Yet spacecraft imagery presents a structural mismatch for uniform 2DGS regression: the target occupies a small, densely structured region within a vast, featureless deep-space background, so a network that minimizes average reconstruction loss inevitably over-invests capacity in the irrelevant background and smears the fine edges of antennas and solar panels. We propose S3R-GS, a saliency-guided framework that embeds semantic spatial priors into the 2DGS pipeline at three levels: an encoder-level module that suppresses background noise before it reaches the splatting stage; a discrete Gaussian routing mechanism that assigns each spatial location to a semantically appropriate kernel group and reformulates Gaussian modeling as semantic prototype selection; and a saliency-weighted training strategy that concentrates the optimization gradient on the spacecraft target. Experiments on the SPEED and SPEED+ benchmarks show that S3R-GS achieves strong PSNR performance, competitive SSIM, and improved perceptual quality across scale factors from × 2 to × 12 ; additional ablation, extreme-lighting, and efficiency analyses further support the robustness and practicality of the proposed design. [ABSTRACT FROM AUTHOR] |
| Copyright of Remote Sensing is the property of MDPI and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Highlights: What are the main findings? This study proposes S3R-GS, an arbitrary-scale super-resolution model that reconstructs spacecraft images at any magnification factor using a single network. The model uses spatial clustering to assign Gaussian kernels strictly to the spacecraft body, avoiding wasteful computations on the featureless deep-space background. What are the implications of the main finding? Relying on explicit 2D Gaussian shapes rather than point-wise regression allows S3R-GS to maintain sharp details even when extrapolated far beyond typical scales (up to ×12). Tests across different datasets demonstrate that stripping away background noise leads to much cleaner object boundaries, which directly benefits subsequent tasks like optical pose estimation. High-resolution images of non-cooperative spacecraft are essential for on-board autonomous operations. Hardware bandwidth limits and continuously changing observation distances mean that a practical super-resolution (SR) system must handle arbitrary, non-integer magnification factors without retraining, a setting known as arbitrary-scale SR (ASSR). Recent 2D Gaussian splatting (2DGS) methods represent image content with explicit anisotropic Gaussian primitives and render at any continuous coordinate, offering substantially faster inference than implicit neural representation (INR) approaches. Yet spacecraft imagery presents a structural mismatch for uniform 2DGS regression: the target occupies a small, densely structured region within a vast, featureless deep-space background, so a network that minimizes average reconstruction loss inevitably over-invests capacity in the irrelevant background and smears the fine edges of antennas and solar panels. We propose S3R-GS, a saliency-guided framework that embeds semantic spatial priors into the 2DGS pipeline at three levels: an encoder-level module that suppresses background noise before it reaches the splatting stage; a discrete Gaussian routing mechanism that assigns each spatial location to a semantically appropriate kernel group and reformulates Gaussian modeling as semantic prototype selection; and a saliency-weighted training strategy that concentrates the optimization gradient on the spacecraft target. Experiments on the SPEED and SPEED+ benchmarks show that S3R-GS achieves strong PSNR performance, competitive SSIM, and improved perceptual quality across scale factors from × 2 to × 12 ; additional ablation, extreme-lighting, and efficiency analyses further support the robustness and practicality of the proposed design. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 20724292 |
| DOI: | 10.3390/rs18101585 |
