2D imaging tools for harmonisation in plastic pollution data

Small 'mesoplastics' and 'microplastics' make up the most abundant visible plastics in the ocean, however, due to diverse research methods among studies on plastic across different marine environments, data on the physical characteristics of these plastics, including size, shape, and colour cannot be directly compared. The use of 2D imaging or image segmentation tools has seen a growth of adaptations to the plastics space in recent years. We comprehensively assesses four methodologies to count and categorise visible meso and microplastics: (1) traditional hand counts, (2) ImageJ, (3) Saturna Imaging System from Ocean Diagnostics, (4) Image Segmentation Model from The Ocean Cleanup, to demonstrate how useful these methods are in harmonising data on the physical characteristics of plastics. We used 2D imaging to characterise plastics collected across three different sampling environments each with different standard methods for reporting colour and type; (a) coastlines (beach cleanups in Tasmania), (b) the sea-surface (manta trawls from the North Pacific Gyre), and (c) within biota (seabird necropsies from the east coast of Australia). We provide a comprehensive assessment of 2D imaging tools, their success in application across the fields of plastic pollution research, and recommendations for reporting 2D imaging outputs. Also see: https://micro2024.sciencesconf.org/559626/document