Deep Learning for Ship Classification on Medium Resolution SAR Imagery
International audience This research delves into the classification of maritime vessels, utilizing medium-resolution Synthetic Aperture Radar (SAR) imagery obtained from Sentinel-1, alongside Automatic Identification System (AIS) data streams. The investigation is specifically designed to address a...
Main Authors: | , , , |
---|---|
Other Authors: | , , , , , , , , , , , , , |
Format: | Conference Object |
Language: | English |
Published: |
HAL CCSD
2023
|
Subjects: | |
Online Access: | https://hal.science/hal-04277648 https://hal.science/hal-04277648v2/document https://hal.science/hal-04277648v2/file/Ship_classification_v12%20%281%29.pdf |
Summary: | International audience This research delves into the classification of maritime vessels, utilizing medium-resolution Synthetic Aperture Radar (SAR) imagery obtained from Sentinel-1, alongside Automatic Identification System (AIS) data streams. The investigation is specifically designed to address a ternary classification challenge involving three distinct ship categories: Tanker, Cargo, and Others. Leveraging a dataset comprising over 80,000 ship images, a Convolutional Neural Network (CNN) ensemble is applied. The results reveal a total classification accuracy of 79%. |
---|