Optical/SAR data and System Integration for Rush Identification of Ship models
Funded under: European Space Agency
Start date 29 March 2021 End date 28 September 2022
Keywords: Maritime traffic surveillance; Optical/SAR image analysis; Ship classification; Ship kinematics estimation; Ship behavior analysis
OSIRIS-FO is the follow-on of ESA's GSTP Project OSIRIS (Optical/SAR data and system Integration for Rush Identification of Ship models, 2016-2018, https://wiki.services.eoportal.org/tiki-index.php?page=OSIRIS). The main objective is to develop, implement and integrate tools for Earth Observation Maritime Surveillance into an Open-Source platform treating both optical and synthetic aperture radar (SAR) satellite imagery and Automatic Identification System (AIS) data from collaborative vessels. The main modules integrated into the final OSIRIS system are the following:
- SD - Ship Detection
- SC - Ship Classification
- SKE - Ship Kinematic Estimation
- SBA - Ship Behavior Analysis
- SRP - Ship Route Prediction
- WG - Web Geographic Information System
- CP - Control Panel
The partnership is composed of:
- Mapsat Srl, project leader, an Italian space company, developing the system for ship detection;
- Sistemi Territoriali Srl, a CNR spinoff IT-consulting company, developing the WebGIS and the CP modules;
- CNR, with the IIT and ISTI institutes, developing ship classification and kinematic estimation (ISTI), and ship route prediction and behavior analysis (IIT).
This follow-on project aims to improve the OSIRIS modules' functionalities for SAR image processing to bring their Technology Readiness Level (TRL) from 6 to 8 and make them support to ESA's Copernicus mission exploitation.
The Signals and Images Lab contribution consists of improvements in the SC module for ship classification from moderate-resolution SAR data, and in the SKE module for reliable vessel velocity estimation by Fourier analysis of single-look complex (SLC) SAR images. All the studies and developments will be carried out taking the ESA's Sentinel-1 image features as the reference.