A marine target extraction model for marine algal bloom extraction with an integration-enhanced adaptive neural algorithm

Marine algal blooms (MABs) have become increasingly problematic on a global scale, representing a significant threat to human health, ecological safety, and socio-economic development. The development of high-precision and robust techniques for extracting MABs from multi-source remote sensing imager...

Full description

Saved in:

Bibliographic Details
Published in:	Expert systems with applications Vol. 271; p. 126541
Main Authors:	Liao, Siyuan, Han, Shuzong, Huang, Haoen, Wang, Daoru
Format:	Journal Article
Language:	English
Published:	Elsevier Ltd 01.05.2025
Subjects:	Integration-enhanced adaptive gradient descent (IEAGD) neural algorithm Marine algal blooms (MABs) Marine target extraction (MTE) model Satellite remote sensing image Integration-enhanced adaptive gradient descent (IEAGD) neural algorithm Satellite remote sensing image Marine target extraction (MTE) model Marine algal blooms (MABs)
ISSN:	0957-4174
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Marine algal blooms (MABs) have become increasingly problematic on a global scale, representing a significant threat to human health, ecological safety, and socio-economic development. The development of high-precision and robust techniques for extracting MABs from multi-source remote sensing imagery has thus far proven to be a formidable challenge. Traditional extraction models and algorithms have been proven to be incapable of adapting to the complex marine environments. To address these limitations, an integral-enhanced adaptive gradient descent (IEAGD) neural algorithm is proposed in this paper. This framework involves automatic adjustment of the step size to facilitate entry into the global optimal solution, enabled by error calculation in previous iterations. Additionally, an error integration term, grounded in control theory, is incorporated to enhance noise tolerance. The IEAGD neural algorithm is analyzed for convergence and proven to be robust. Finally, the superiority of the proposed IEAGD neural algorithm is confirmed by several comparison experiments using different remote sensing images of marine algal blooms in various noisy working environments. •Constructed MTE model can be used effectively in complex marine environments.•IEAGD algorithm is robust to various noises by the integration-enhanced term.•Experiment resutls show the robustness and superior extraction performance of IEAGD algorithm.
ISSN:	0957-4174
DOI:	10.1016/j.eswa.2025.126541