Review of current perspectives and emerging trends in ship wake detection

Authors:Irene Ortiz Abuja, Luis Ciruelo Sanz, Enrique Diez Benito, Francisco Rodríguez Robles, Víctor Chaparro Parra, Manuel Naranjo Martínez y Roberto Gómez-Espinosa Martín.

Abstract

This paper explores the field of ship wake detection using satellite imagery. The research examines traditional methods, such as the Radon transform and the Hough transform, as well as more modern techniques based on deep learning. The authors discuss methods for estimating parameters such as ship speed and direction from wake features, and discuss the importance of high-quality labeled data sets for the development of machine learning models.

Implementation of a Deep Learning Model for the Detection of vessels alongside in Satellite Imagery in SEDA

Authors:Manuel Naranjo Martínez, Luis Ciruelo Sanz, Jose María Moreu Gamazo, Enrique Diez Benito, Víctor Chaparro Parra, y Roberto Gómez-Espinosa Martín.

Abstract

This paper explores the detection of vessels alongside in satellite imagery, an act that may be associated with illegal activities. Instead of relying on the Automatic Identification System (AIS), which can be disabled to avoid detection, deep learning is used to analyze the imagery. Two models are presented: one for detecting vessels and one for classifying vessels by identifying those that approach other vessels.

State of the Art Review of Artificial Intelligence Algorithms on Satellite Imagery for Navigation Route Planning in Polar Zones

Authors: Francisco Rodríguez Robles, Guillermo Rodríguez Llorente, Gustavo Aguado Perianes, Enrique Diez Benito, Irene Ortiz Abuja, Manuel Naranjo Martínez y Roberto Gomez-Espinosa

Abstract

This paper presents a review of the state of the art in the use of artificial intelligence (AI) algorithms to analyze satellite imagery in support of navigational route planning in polar areas. The paper highlights the growing commercial, tourism and military interest in polar regions and the need for advanced routing systems that minimize cargo damage and ensure timely arrivals. The paper focuses on two main approaches: drift ice detection and sea ice concentration estimation.

The most important papers comparing or presenting different alternatives related to processing techniques and model architectures are analyzed, applying the conclusions drawn to the development of the MADS (MAduration of the SEDA technology Demonstrator) project, which seeks to evolve a geospatial intelligence platform of the Spanish Ministry of Defense called SEDA (SatEllite Data AI).

Implementation of a system based on Deep Learning for the detection of dark vessels in satellite images through route prediction.

Authors: Enrique Diez Benito, Jose María Moreu Gamazo, Luis Ciruelo Sanz, Víctor Chaparro Parra, Manuel Naranjo Martínez y Roberto Gómez-Espinosa Martín.

Abstract

This study presents a system for detecting “dark vessels” (ships that turn off their Automatic Identification System -AIS-) using artificial intelligence and satellite images. The system uses a deep learning model to predict the route of a ship after turning off its AIS, and subsequently detects it in a satellite image. The model is based on historical AIS data and is applied to Sentinel-2 imagery to pinpoint possible locations of the ship. This system can be useful for improving maritime security, crime fighting and suspicious ship detection.