VE_ESA-TRACE (DTA.AD005.320)
Thematic area
Earth system science and environmental technologies
Project area
Tecnologie e processi per l'ambiente (DTA.AD005)Structure responsible for the research project
Marine science institute (ISMAR)
Project manager
MICHOL GHEZZO
Phone number: +39 041 2407941
Email: michol.ghezzo@ve.ismar.cnr.it
Abstract
The overall goal is to obtain precise and reliable data on floating macro-litter regarding their quantity, position,
accumulation zones, material properties, floating depth, and sources, which may serve as a basis for litter recovery,
source elimination, and prevention of litter dispersal. The combination of modern satellite technology, deep learning
and trajectories forecast can potentially be applied wordwilde and can help to monitor the open sea.
The project will refine exisiting methods to detect the marine litter starting by earth observation data. The trajectories
of the marine litter will be also calculated by a forecast circulation and particle tracking numerical model. The
validation of the system and the GIS analysis of the results will produce a powerful relocable tools, available by a web
interface to support potentially the stackeholders in the marine litter assessment, monitoring and cleaning.
Goals
Our goal is to leverage the full potential of deep learning (DL) and of the synergies of combining optical and SAR
remote sensing with floating litter trajectory forecasting as calculated by the oceanographic forecasting system, hence
creating a technical solution that can beused in practice to support cleanup actions.
The proposed system will be easily relocable in regional seas by simply exchanging the regional ocean current model
component.
The tracked objects can potentially be used to substantially improve regional ocean current models worldwide through
data assimilation and time series comparison. The interaction between remote sensing detection and trajectory
forecast modeling will solve the crucial problem of initial displacement in marine litter dispersal simulation.
The system can help scientific institutions, governmental authorities and political decision-makers to fill knowledge
gaps for areas where little is known about litter pollution. The collected data and most importantly the trajectory
forecasting component with its web interface can support cleanup missions as well as regional initiatives.
Start date of activity
01/08/2020
Keywords
Macro-Litter, Remote Sensing, Machine Learning
Last update: 10/07/2025