Journal
MARINE TECHNOLOGY SOCIETY JOURNAL
Volume 55, Issue 1, Pages 38-49Publisher
MARINE TECHNOLOGY SOC INC
Keywords
unmanned aerial vehicles; computer vision; vessel detection; tracking; identification
Categories
Funding
- POFC (Programa Operacional Factores de Competitividade) within the National Strategic Reference Framework (QREN) [2013/034063, 34063]
Ask authors/readers for more resources
The paper discusses the development of an integrated system for maritime situation awareness using unmanned aerial vehicles, with a focus on the role of the automatic detection subsystem. Utilizing sensors onboard UAVs for automatic vessel detection, the system aims to assist human operators in generating situational awareness of maritime events effectively and efficiently. The system demonstrates high precision rates, low latency, and suitable recalls, comparable to other state-of-the-art approaches, through field tests detecting multiple vessels and lifesavers in different types of images.
This paper addresses the development of an integrated system to support maritime situation awareness based on unmanned aerial vehicles (UAVs), emphasizing the role of the automatic detection subsystem. One of the main topics of research in the SEAGULL project was the automatic detection of sea vessels from sensors onboard the UAV, to help human operators in the generation of situational awareness of maritime events such as (a) detection and geo-referencing of oil spills or hazardous and noxious substances, (b) tracking systems (e.g., vessels, shipwrecks, lifeboats, debris), (c) recognizing behavioral patterns (e.g., vessels rendezvous, high-speed vessels, atypical patterns of navigation), and (d) monitoring environmental parameters and indicators. We describe a system composed of optical sensors, an embedded computer, communication systems, and a vessel detection algorithm that can run in real time in the embedded UAV hardware and provide to human operators vessel detections with low latency, high precision rates (about 99%), and suitable recalls (>50%), which is comparable to other more computationally intensive state-of-the-art approaches. Field test results, including the detection of lifesavers and multiple vessels in red-green- and-blue (RGB) and thermal images, are presented and discussed.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available