Plastic Litter Project: Coastal Litter Mapping

In a relatively short period, the attributes of plastic initially perceived to be positive characteristics - convenience and longevity - have shifted to pose a widespread environmental problem. Within the marine context, millions of tonnes of plastic enter our oceans annually. The economic cost to marine natural capital alone is estimated to range from $3300–$33,000 per ton of plastic per year.

In this context, our project has a decisive contribution to the design and coordination for relevant actions to monitor natural and human systems interaction, exploit beneficial opportunities, and monitor anthropogenic marine litter accumulation.

Here is where our group Marine Remote Sensing Group MRSG and Plastic Litter Project PLP comes in.

The answer can be found at our coastlines. The MRSG team is using UAVs (Unmanned Aerial Vehicle) to acquire images from a series of coastal zones with high-resolution optical sensors. Using the power of deep learning algorithms, the team trains AI algorithms that can map and quantify marine litter washing onto our beaches.

Where all of you can help?
Here is where we need YOUR valuable help to classify images contain marine litter. You will help us create a dataset for our deep learning algorithms and help our team to map marine litter accumulation. The more images you classify, the better the machine learning algorithms get.

Aims & Goals
The project aims are:

1. to train our deep learning algorithms from YOUR Zooniverse classifications,
2. to create marine litter accumulation maps.

Geographic area of the project
At the moment the geographic area of the project is the Aegean Sea in Greece.
But our dream is to expand the geographic coverage and detect marine litter in the whole world!

Results
Density maps created from classification results can be found at Coastal Marine Litter Observatory by MRSG

The PLP2020 project
This project is an extension of the Plastic Litter Project 2020 (PLP2020) conducted by the MRSG team. PLP2020 is a project targeting on detecting and validating artificial plastic targets on the sea surface using UAV and satellite image technology.

Publications
The results will be presented in the future at conferences and papers.
Our last publication in Drones Journal can be found at:
A Citizen Science Unmanned Aerial System Data Acquisition Protocol and Deep Learning Techniques for the Automatic Detection and Mapping of Marine Litter Concentrations in the Coastal Zone