Abstract
Most data collected with autonomous underwater vehicles (AUVs) can only be reviewed and analyzed post-mission. This limitation is due to the low-bandwidth acoustic links used for communication between AUV and surface, which are typically devoted to monitoring a few mission-critical parameters. In this work, we propose a plug-and-play solution that would enable live streaming of side-scan sonar data on such low-bandwidth modems with no additional hardware requirements. Using principles from image super resolution, we first downsample and compress sonar images on the sensor(AUV)-side and then estimate the full high-quality sonar image on the receiver(surface)-side. The bandwidth requirements for this livestreaming are below 1.87 kbit/s while still providing reconstructions difficult to distinguish from the originals. Our complete end-to-end system has been implemented using EvoLogics underwater modems and successfully tested in our laboratory and harbor basin.
| Original language | English |
|---|---|
| Book series | IFAC-PapersOnLine |
| Volume | 54 |
| Issue number | 16 |
| Pages (from-to) | 458-463 |
| ISSN | 2405-8963 |
| DOIs | |
| Publication status | Published - 2021 |
Keywords
- Acoustic data transmission
- Autonomous underwater vehicle (AUV)
- Deep learning
- Side-scan sonar imaging
- Super-resolution
Fingerprint
Dive into the research topics of 'Side-scan sonar imaging: Real-time acoustic streaming'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver