REGISTER NOW AND BOOK HOTEL

Tutorials

OCEANS 2025 is offering tutorials for attendees with the purpose of expanding the attendees’ knowledge of key technologies or emerging areas of technology and research. The tutorials will take place on Monday, September 29 in Room 305.

With speakers/instructors providing rare insights into the subject and covering both theoretical & practical aspects in sufficient depth, tutorials are expected to benefit academia and industry. The list of tutorials proposed to be conducted subject to enough interest from registering attendees, is given below.

Meals or refreshments will not be provided for tutorial attendees – you are required to make your own arrangements for this during the tea/lunch breaks. Tutorial materials will be made available to attendees before the beginning of the conference.

List of Tutorials

Presenter(s):

Katie Skinner, Assistant Professor, Robotics Department, University of Michigan
Onur Bagoren, PhD Candidate, Robotics Department, University of Michigan
Anja Sheppard, PhD Candidate, Robotics Department, University of Michigan

Length: 2 hours

Time: 12:30-2:30 PM

Room: 305

Description:

Recent advances in sonar technology and marine robotic systems have enabled efficient, large area data collection in oceans and the Great Lakes. This has resulted in massive amounts of data available for marine science and engineering applications. Artificial intelligence (AI) and machine learning (ML) have potential to automate data processing and accelerate discovery from sonar data. This tutorial will focus on AI/ML for automated shipwreck detection. We will present the AI4Shipwrecks dataset, a new side scan sonar dataset featuring 28 distinct shipwreck sites in Thunder Bay National Marine Sanctuary in Lake Huron, MI. We will also provide a software demonstration of a new QGIS tool for automated shipwreck detection from multibeam sonar data. 

Tentative Schedule:

  • Introduction
  • Presentation on AI4Shipwrecks dataset for shipwreck detection from side scan sonar data
  • Demo/tutorial of QGIS tool for shipwreck detection from multibeam sonar data
  • Q&A and Open Discussion

Figure. This tutorial presents AI/ML methods for shipwreck detection from sonar data.

Funding Acknowledgement: This work was supported by grants from NOAA Ocean Exploration under awards #NA21OAR0110196 and #NA23OAR0110315.

Other Links:

AI4Shipwrecks Dataset: https://umfieldrobotics.github.io/ai4shipwrecks/
NOAA Ocean Exploration Expedition Description: https://oceanexplorer.noaa.gov/explorations/22shipwreck-detection/welcome.html 
“Building curious machines” by Michigan Engineering: https://news.engin.umich.edu/2023/05/building-curious-machines/