Plastic is convenient. It's cheap, useful, and it can be moulded into very amusing shapes. The trouble is that it never goes away. While plastic is recyclable and can be reused as lesser quality plastic, it can never be completely eliminated. Instead, it breaks down into smaller and smaller fragments, called microplastics (MPs), that persist in the marine environment in view of their obdurate nature.
There is widespread recognition that monitoring of MPs is crucial due to the possible uptake by organisms at low trophic levels. So much so, that one of the indicators currently entrenched within the Marine Strategy Framework Directive (MSFD) refers to trends in the amount, distribution, and where possible, composition of micro-particles within the marine environment.
Despite this consensus in recognising the importance of studying MPs, the routine protocols for monitoring their abundance are time-consuming and involve the examination of samples with the naked eye or with a microscope. This creates bias in the recorded subjective attributes such as colour and roughness since, withinsuch visual analysis, human observer interpretation is inevitable. For instance, different observers may assign different colours when carrying out the analysis under different prevailing lighting conditions.
Dr Adam Gauci, Prof. Alan Deidun, and Mr John Montebello from the Physical Oceanography Research Group of the Department of Geosciences, together with Dr John Abela from the Department of Computer Information Systems within the Faculty of ICT, and Prof. François Galgani from the Institut Français de Recherche pour l'Exploitation de la Mer (IFREMER), are investigating the use of image processing techniques to automatically extract the required parameters from digitised samples.
Apart from making the process faster, an automated method also removes the subjectivity and allows more accurate spatial and temporal comparisons to be made. The proposed algorithms have already been tested on a large number of MP samples collected over several months from four different beaches along the Maltese coastline by applying the same algorithm to high-resolution scans of well-spaced individual MP particles. Exploratory correlations with attributes quantified by a human observer, were also carried out for validation purposes and the results confirm that the developed algorithms perform well.
This work has been accepted for publication in the special 'Curbing the Inexorable Rising on Marine Litter' issue of the 'Ocean Coastal and Management' journal. The suggested sampling protocols and developed image processing techniques will also be presented during the upcoming 2019 CIESM Congress being held in Cascais, Lisbon, between 7 and 11 October. The development and validation of an image analysis MP characterisation algorithm paves the way for a broader future application of the same algorithm, involving a possible incorporation within a smart phone app, for instance.
