Skip to main content

IMIS

A new integrated search interface will become available in the next phase of marineinfo.org.
For the time being, please use IMIS to search available data

 

[ report an error in this record ]basket (0): add | show Print this page

CNN-based object detection and segmentation for maritime domain awareness
Nita, C.; Vandewal, M. (2020). CNN-based object detection and segmentation for maritime domain awareness, in: Dijk, J. (Ed.) Artificial Intelligence and Machine Learning in Defense Applications II. Proceedings of SPIE, the International Society for Optical Engineering, 11543: pp. 1154306. https://hdl.handle.net/10.1117/12.2573287
In: Dijk, J. (Ed.) (2020). Artificial Intelligence and Machine Learning in Defense Applications II. Proceedings of SPIE, the International Society for Optical Engineering, 11543. SPIE: Washington. ISBN 9781510638990; e-ISBN 9781510639003. , more
In: Proceedings of SPIE, the International Society for Optical Engineering. SPIE: Bellingham, WA. ISSN 0277-786X; e-ISSN 1996-756X, more
Peer reviewed article  

Available in  Authors 
Document type: Conference paper

Keyword
    Marine/Coastal
Author keywords
    Ship intelligence; deep neural network; vessel detection; image segmentation; maritime domain awareness

Authors  Top 

Abstract
    Deep learning algorithms have been proven to be a powerful tool in image and video processing for security and surveillance operations. In a maritime environment, the fusion of electro-optical sensor data with human intelligence plays an important role to counter the security issues. For instance, the situational awareness can be enhanced through an automated system that generates reports on ship identity and signature together with detecting the changes on naval vessels activity. To date, various studies have been set out to explore the performance of deep neural networks using a ship signature database. In the current study, we investigate the Mask R-CNN method to address not only the naval vessel detection using bounding boxes, but also obtaining their segmentation masks. We train and validate the model on data captured by an on-board camera covering the visible spectral band under various weather and light conditions. The experimental results show that Mask R-CNN provides high confidence scores on challenging scenarios with a mean average precision of 86.4%. However, the precision of the segmentation mask is slightly deteriorated when the ships are adjacent to the border of the captured scene. Moreover, the network tested on thermal images indicates a decrease in detection and segmentation performance since the training data distribution is not representative enough.

All data in the Integrated Marine Information System (IMIS) is subject to the VLIZ privacy policy Top | Authors