Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Recent Advances in Molecular Engineering of Organelle-Targeted Viscosity Probes for Fluorescence and Biological Applications.

ACS sensors·2026
Same author

Laparoscopic versus open pancreaticoduodenectomy for non-pancreatic periampullary adenocarcinomas: study protocol for a multicentre randomised controlled trial in China.

BMJ open·2026
Same author

Optimizing nitrogen management enhances photosynthesis and alfalfa productivity in jujube-alfalfa intercropping.

Scientific reports·2026
Same author

Diagnostic Performance of Serum AKR1B10 in Early-Stage and AFP-Negative Hepatocellular Carcinoma: A Multicentre Study.

Liver international : official journal of the International Association for the Study of the Liver·2026
Same author

Evidence supporting the role of GIGYF2 in synapse development and autism.

Molecular psychiatry·2026
Same author

Fusion prediction model for post-ERCP pancreatitis under NSAIDs prophylaxis.

Surgical endoscopy·2026

Related Experiment Video

Updated: Sep 9, 2025

Long-term Behavioral Tracking of Freely Swimming Weakly Electric Fish
10:56

Long-term Behavioral Tracking of Freely Swimming Weakly Electric Fish

Published on: March 6, 2014

12.6K

Lightweight underwater debris detection model based on improved RT-DETR.

Conggong Lin1, Yushi Zhang1, Guodong Chen1

  • 1College of Physics and Information Engineering, Fuzhou University, Fuzhou, 350108, China.

Marine Pollution Bulletin
|August 29, 2025
PubMed
Summary

Researchers developed LCSA-DETR, a lightweight object detection model for autonomous underwater vehicles (AUVs) to monitor marine debris. This efficient model achieves high accuracy while significantly reducing computational demands for real-time underwater applications.

Keywords:
Lightweight networkRT-DETRStarNetUnderwater garbage

More Related Videos

Reefshape: A System for the Efficient Collection and Automated Processing of Time-Series Underwater Photogrammetry Data for Benthic Habitat Monitoring
13:35

Reefshape: A System for the Efficient Collection and Automated Processing of Time-Series Underwater Photogrammetry Data for Benthic Habitat Monitoring

Published on: June 13, 2025

638
A Field Primer for Monitoring Benthic Ecosystems Using Structure-From-Motion Photogrammetry
06:36

A Field Primer for Monitoring Benthic Ecosystems Using Structure-From-Motion Photogrammetry

Published on: April 15, 2021

3.8K

Related Experiment Videos

Last Updated: Sep 9, 2025

Long-term Behavioral Tracking of Freely Swimming Weakly Electric Fish
10:56

Long-term Behavioral Tracking of Freely Swimming Weakly Electric Fish

Published on: March 6, 2014

12.6K
Reefshape: A System for the Efficient Collection and Automated Processing of Time-Series Underwater Photogrammetry Data for Benthic Habitat Monitoring
13:35

Reefshape: A System for the Efficient Collection and Automated Processing of Time-Series Underwater Photogrammetry Data for Benthic Habitat Monitoring

Published on: June 13, 2025

638
A Field Primer for Monitoring Benthic Ecosystems Using Structure-From-Motion Photogrammetry
06:36

A Field Primer for Monitoring Benthic Ecosystems Using Structure-From-Motion Photogrammetry

Published on: April 15, 2021

3.8K

Area of Science:

  • Marine robotics
  • Computer vision
  • Environmental monitoring

Background:

  • Underwater ecosystems are threatened by anthropogenic debris.
  • Autonomous monitoring and removal are crucial for addressing these threats.
  • Resource-constrained autonomous underwater vehicles (AUVs) require efficient object detection models.

Purpose of the Study:

  • To develop a lightweight object detection model (LCSA-DETR) optimized for AUVs.
  • To improve detection performance and efficiency for underwater debris monitoring.
  • To enable real-time deployment on resource-constrained platforms.

Main Methods:

  • Introduced LCSA-DETR, based on RT-DETR, with architectural modifications: StarNet backbone, lightweight cross-stage aggregation encoder (LC-Encoder), adaptive kernel fusion block (AKFB), and bidirectional feature pyramid network (BiFPN).
  • Applied layer-adaptive magnitude-based pruning (LAMP) to create a compressed version, LCSA-DETR-P.
  • Evaluated performance on the Trash-ICRA19 dataset without pre-trained weights.

Main Results:

  • LCSA-DETR-P achieved 80.4% AP and 98.3% AP50 on the Trash-ICRA19 dataset.
  • The model significantly reduced parameters (31.7%), FLOPs (23.0%), and model size (35.0%) compared to the baseline.
  • Achieved a real-time inference speed of 105.1 frames per second.

Conclusions:

  • LCSA-DETR-P offers accuracy comparable to RT-DETR-R18 with substantially lower computational requirements.
  • The model's efficiency and real-time performance make it suitable for deployment on resource-constrained AUVs for underwater debris detection.
  • This work contributes to advancing autonomous solutions for marine debris monitoring and management.