About me
Alexandre Cardaillac is a Postdoctoral Research Associate in Underwater Computational Imaging at the University of Sydney. He is part of the ARIAM Research Hub, the Australian Centre for Robotics, and the Robotic Imaging Lab. His main research interests include underwater computational imaging using acoustic and optical data, scene understanding, and situation awareness for underwater robotic systems with the aim the develop further the autonomy and safety of such vehicles.
Alexandre received a Bachelor of Information Technology from the Nantes School of Digital Innovation in 2019 and his M.Sc. degree in Artificial Intelligence with Speech and Multimodal Interaction from the Heriot-Watt University in 2020 and received the award of the best M.Sc. dissertation for his work on uncertainty estimation in deep neural networks. He completed his Ph.D. degree in Engineering in 2023 with the Department of Marine Technology at the Norwegian University of Science and Technology, as part of the Applied Underwater Robotics Laboratory.
Selected publications
ROV-Based Autonomous Maneuvering for Ship Hull Inspection with Coverage Monitoring
Journal of Intelligent & Robotic Systems, 2024Alexandre Cardaillac, Roger Skjetne and Martin Ludvigsen [paper]
The proposed work aims at developing the basis for an end-to-end autonomous solution for ship hull inspection. A sonar is employed to perform hull-relative navigation at the same time as incrementally build an inspection map for the operator to monitor the operation.
- Adaptive coverage path planning leveraging sensor coverage capabilities.
- Sonar-based maneuvering-based guidance strategy for hull relative navigation.
- Real-time generation of an acoustic inspection map for inspection monitoring.
The proposed methods were extensively tested on ships of different sizes and shapes and in multiple harbours across Europe. The results were validated by domain experts which makes the proposed work industry-relevent.
Camera-Sonar Combination for Improved Underwater Localization and Mapping
IEEE Access, 2023Alexandre Cardaillac and Martin Ludvigsen [paper]
This article aims at combining a monocular camera and a forward looking sonar for improved underwater localisation and mapping. The advantages of both sensors are leveraged to compensate for their respective limitations and constraints.
- Feature-based visual-acoustic combination (monocular camera + sonar).
- Continuous, stable, and robust trajectory and point cloud rescaling.
- Real-time performance and SLAM integration.
The proposed approach was tested in underwater inspection scenarios, including propeller and subsea structure inspections, and outperformed alternative solutions.
Semantic Segmentation in Underwater Ship Inspections: Benchmark and Data Set
IEEE Journal of Oceanic Engineering, 2022Maryna Waszak, Alexandre Cardaillac, Brian Elvesæter, Frode Rødølen and Martin Ludvigsen [paper]
In this article, we present the first large-scale data set for underwater ship inspection. It is designed for detection of ship parts and faults.
- 1893 images with segmentation masks for 10 categories (6 ship parts + 4 faults).
- Dataset analysis and benchmarks.
- Reviewed and validated by domain experts.
We believe the proposed dataset create new promising opportunities for future research in underwater ship inspections. It is made available for noncommercial use on https://liaci.sintef.cloud.
Publication list
- Towards autonomous underwater navigation and perception for end-to-end ship hull inspection, PhD Thesis, 2024, Alexandre Cardaillac [thesis]
- ROV-Based Autonomous Maneuvering for Ship Hull Inspection with Coverage Monitoring, Journal of Intelligent & Robotic Systems, 2024, Alexandre Cardaillac, Roger Skjetne and Martin Ludvigsen [paper]
- Camera-Sonar Combination for Improved Underwater Localization and Mapping, IEEE Access, 2023, Alexandre Cardaillac and Martin Ludvigsen [paper]
- Application of Maneuvering Based Control for Autonomous Inspection of Aquaculture Net Pens, 8th Asia-Pacific Conference on Intelligent Robot Systems (ACIRS), 2023, Alexandre Cardaillac, Herman Biørn Amundsen, Eleni Kelasidi and Martin Ludvigsen [paper]
- Fusion of Multi-Modal Underwater Ship Inspection Data with Knowledge Graphs, OCEANS 2022, Hampton Roads, 2022, Joseph Hirsch, Brian Elvesæter, Alexandre Cardaillac, Bernhard Bauer and Maryna Waszak [paper]
- Modular Multi-Sensor Fusion for Underwater Localization for Autonomous ROV Operations, OCEANS 2022, Hampton Roads, 2022, Martin Scheiber, Alexandre Cardaillac, Christian Brommer, Stephan Weiss and Martin Ludvigsen [paper]
- Marine Snow Detection for Real Time Feature Detection, 2022 IEEE/OES Autonomous Underwater Vehicles Symposium (AUV), 2022, Alexandre Cardaillac and Martin Ludvigsen [paper]
- A Communication Interface for Multilayer Cloud Computing Architecture for Low Cost Underwater Vehicles, 11th IFAC Symposium on Intelligent Autonomous Vehicles IAV, 2022, Alexandre Cardaillac and Martin Ludvigsen [paper]
- Semantic Segmentation in Underwater Ship Inspections: Benchmark and Data Set, IEEE Journal of Oceanic Engineering, 2022, Maryna Waszak, Alexandre Cardaillac, Brian Elvesæter, Frode Rødølen and Martin Ludvigsen [paper]
- Path Following for Underwater Inspection Allowing Manoeuvring Constraints, Intelligent Autonomous Systems 17, 2022, Alexandre Cardaillac and Martin Ludvigsen [paper]
- Ruled Path Planning Framework for Safe and Dynamic Navigation, OCEANS 2021: San Diego - Porto, 2021, Alexandre Cardaillac and Martin Ludvigsen [paper]