Detection and trajectory extraction of streak-like objects in wide-field astronomical images with slice-assisted deep learning
Clicks: 1
ID: 314513
2026
Article Quality & Performance Metrics
Overall Quality
Not rated
Combines reader engagement with the AI quality analysis. This
article has not been analysed, so there is no overall score —
reader engagement is measured and shown alongside.
Reader Engagement
0.0
/100
1 views
0 readers
AI Quality Assessment
Not analyzed
Readership in this journal
Ranked #517 of 900 articles by views in monthly notices of the royal astronomical society
Most read
Least read
Bar heights use a square-root scale. Only the 120 most-read articles are drawn; the journal has 900 in total.
Mint this article as an NFT
Not yet mintedCreate a permanent, verifiable on-chain record of this article on the Scimatic Network. The NFT is held in your Journament account, and you can withdraw it to your own wallet at any time.
5
SUSD
one-off · no wallet required
Abstract
Abstract Wide-field optical surveys play a significant role in modern astronomy, supporting diverse scientific applications from time-domain astrophysics to the monitoring of near-Earth objects and man-made satellites. In sidereal tracking mode, moving objects frequently appear as elongated streaks. Detecting these linear structures is critical for efficient data reduction, however, their reliable detection and association pose significant challenges due to dense stellar backgrounds and variable observing conditions. We present an automated pipeline for the detection and trajectory extraction of streak-like moving object images in wide-field astronomical frames. The framework integrates a slice-based deep learning inference for frame-level streak detection with a distance-constrained Hungarian algorithm for trajectory association. Furthermore, a motion-aware refinement module, incorporating velocity estimation, position prediction and confidence-guided re-association is introduced to mitigate the track fragmentations without relying on end-to-end tracking models. Our pipeline is evaluated using the raw images from a trial survey. A total of 2,781 trajectories are extracted, of which 2,543 (approximately 91%) are successfully correlated with the ephemerides, corresponding to 1,359 known objects. Compared to a classical approach our method yields an approximately 11% increase in the number of correlated trajectories and a 14% increase in the number of detected objects, while maintaining the comparable brightness and orbit distributions. The runtime measurements indicate that the per-frame processing times are around several seconds on a standard workstation. These results demonstrate that the proposed framework provides a practical, scalable and solution independent of prior information for streak detection and trajectory extraction in wide-field surveys, supporting future scientific exploration and planetary defense applications.
| Reference Key |
openalex_W7162041001
Use this key to autocite in the manuscript while using
SciMatic Manuscript Manager or Thesis Manager
|
|---|---|
| Authors | He Zhao, 孙荣煜, Shengxian Yu |
| Journal | monthly notices of the royal astronomical society |
| Year | 2026 |
| DOI |
10.1093/mnras/stag958
|
| URL | |
| Keywords | Keywords not found |
Citations
No citations found. To add a citation, contact the admin at info@scimatic.org
Comments
No comments yet. Be the first to comment on this article.