Intelligent airborne monitoring of irregularly shaped man-made objects in the marine ecosystem using statistical Machine Learning techniques

Kuru, Kaya orcid iconORCID: 0000-0002-4279-4166, Clough, Stuart, Ansell, Darren orcid iconORCID: 0000-0003-2818-3315, McCarthy, John and McGovern, Stephanie (2022) Intelligent airborne monitoring of irregularly shaped man-made objects in the marine ecosystem using statistical Machine Learning techniques. IEEE Journal of Oceanic Engineering . ISSN 0364-9059 (Submitted)

Full text not available from this repository.

Official URL: https://papers.ssrn.com/sol3/papers.cfm?abstract_i...

Abstract

The maritime economy has always been diverse and abundant. The habitats of the marine ecosystem, current characteristics of specific types of species and diverse landscape of maritime industry around these habitats are of interest to many researchers, authorities and policymakers. With the applications of emerging fields of science and technology in new and existing industries, prominent companies and research organisations have been recently developing and deploying evolving technologies supported with advanced maritime mechatronics systems (AMMSs) to explore and exploit this tough landscape. This massively evolving industry has the potential to impact the marine ecosystem dramatically; in particular, the seabed, birds, turtles and fish. Automated detection, location and immediate monitoring of these types of industry fields along with the marine life around these areas may be helpful to i) reveal the current impacts, ii) model future possible ecological trends and iii) determine required policies accordingly leading to the reduced ecological footprint and increased sustainability. New automatic techniques are required in order to observe this large environment efficiently. This study aims to develop a novel platform to detect marine man-made objects with irregular shapes in an automated manner in very large datasets composed of maritime aerial survey images. Within this context. a new non-parametric approach has been built, which employs several hybrid statistical Machine Learning (ML) techniques to segment maritime man-made objects (either stationary or mobile) on the sea surface automatically in very large surveys. The validation of the approach has been conducted on a number of aerial maritime domains resulting in successful empirical evidence for its viability. With the approach, man-made objects with irregular shapes without prior training can be detected automatically with the sensitivity and specificity values higher than 0.95. The approach mainly can be employed for several reasons, in particular, will help researchers and policymakers monitor and control diverse maritime industry deployment and dictate required policies to lead on the legal and regulatory requirements to both support maritime technological innovation and shape the future of the maritime industry with respect to the dynamics of the maritime ecosystem.


Repository Staff Only: item control page