A key challenge for maritime security is automatically detecting and classifying ships. This is essential so that government and law enforcement agencies know where ships are and what they are up to. This way they can combat key maritime threats such as piracy, unsustainable fishing, pollution, or the smuggling of illicit goods. One of the main technologies used for maritime security is synthetic aperture radar (SAR) satellite imagery. As a non-cooperative system, SAR can detect ships even if they want to remain hidden. However, SAR has some disadvantages, including low image resolutions and infrequent updates of each location. It is therefore important to understands its limitations regarding commonly confused vessel types and the impact of data quality on classification. The goal is to design a system that can assign confidence scores to detections and classifications, allowing operators to determine when additional data sources are required.

Required Skills:

• Machine learning
• Programming in Python
• Data analysis
• Some fa­mi­li­a­ri­ty with software-defined radio data and signal processing

[1] Fernando Paolo, et al. xview3-sar: Detecting dark fishing activity using synthetic aperture radar imagery. Advances in Neural Information Processing Systems, 35, 2022.

[2] Xiyue Hou, et al. Fusar-ship: Building a high-resolution sar-ais matchup dataset of gaofen-3 for ship detection and recognition. Science China Information Sciences, 63, 2020.

[3] Boying Li, et al. Opensarship 2.0: A large-volume dataset for deeper interpretation of ship targets in sentinel-1 imagery. In SAR in Big Data Era: Models, Methods and Applications. IEEE, 2017.

[4] Chi Zhang et al., Development and Application of Ship Detection and Classification Datasets: A review, in IEEE Geoscience and Remote Sensing Magazine, vol. 12, no. 4, pp. 12-45, 2024.