Observed arrival · 2026-11-05
DeePer: underwater markers for correcting vehicle drift
A bilingual research summary describes a deep-learning method for detecting fiducial markers in murky underwater images to correct uncrewed vehicle localization.
- For
- Underwater robotics and subsea inspection researchers
- Worth noticing
- The comparison links separate per-image counts, differences, confidence intervals, and timing; extracted headline values appear as em dashes.
Field notes
The page explains why marker identity matters: a missed reading leaves accumulated drift, while a wrong ID can relocate the vehicle estimate to another surveyed marker. Its described pipeline locally enhances images with CLAHE before deep-learning detection, and the trial compares DeePer with DeepArUco++ and DeepTag. Four CSV files are linked for per-image counts, differences, confidence intervals, and timing; the extracted headline result fields themselves appear as dashes.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue