Automating the characterisation of beach microplastics through the application of image analyses

Abstract

Four sandy beaches on the island of Malta in the Central Mediterranean were regularly sampled for Large MicroPlastic (LMP) particles having a diameter between 1 mm and 5 mm, at stations located at the waterline and 10 m inshore. The 10975 extracted LMP particles were characterised (dimensions, surface roughness, colour) through unaided visual observation, microscopic analyses, and an algorithm developed within the current study. Two-thirds of the isolated particles were smooth and the majority of these belonged to the grey-white colour category, with a low degree of opaqueness and discolouration, and a high degree of transparency, suggesting that these were dominated by low-density polyethylene LMPs. Conflicting evidence concerning the relative residence time of the isolated LMPs within seawater emerged from the colour and contour roughness determination, although an abundance of primary LMPs (production pellets) within our sample might have been responsible for the low contour roughness results obtained. Roughly six times as many particles were recorded within the inshore sampling stations as the particles recorded at the waterline stations. The developed algorithm performed very well when the dimension and colour parameter values it delivered were compared with those obtained by microscopic analyses. Highlights ► No universal methodology of high validity for analysing isolated microplastics. ► Use of image processing techniques to automatically extract parameters for LMPs. ► Make the process less timeconsuming and removes subjectivity. ► Samples collected from a number of popular beaches around the island of Malta. ► Algorithm performed well in determining the dimensions and colour of the LMPs.

SMACC: A System for Microplastics Automatic Counting and Classification

“…This is mainly due to the demand in terms of time and resources involved in collecting samples and processing them in the laboratory. In sediment studies, the most common identification technique is the visual counting and sorting under a stereomicroscope, which is based on color, size, brightness, and morphology of the particles [9]- [11]. In the case of samples from beaches where the number of particles is high, identifying polymers by Fourier-transform infrared spectroscopy (FTIR) or Raman spectroscopy is a time-consuming task that may exceed the objectives of the study.…”

Results of VAMAS Survey Regarding Microplastic Issues

“…1. assessment of sampling in different matrices (air, waters and leachates, soils and sands, marine organisms, food) to avoid/limit sample contamination during sampling, storage, extraction and detection 57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72 ; 2. definition of reference standard MPs; 3. definition of protocols for dispersion or homogenisation of the microplastics onto the test media 68 ; 4. detection, identification of the chemical components, classification 73,74,75,76,77,78,79,49,68,56,48,80,50,81,52 ;…”

Deep learning approach for automatic microplastics counting and classification

“…The use of stereo microscope for counting and sorting particles based on color, size, brightness, and morphology is the most common identification technique in sediment studies (Hanvey et al, 2017;Gauci et al, 2019) in the range of 1 -5 mm. These parameters are widely used in the study of microplastics (Kooi and Koelmans, 2019).…”

“…Gauci et al (Gauci et al, 2019) proposed the use of Matlab software to analyze the microplastic particles extracted from four beaches in Malta, computing three descriptors for each particle: size, roughness, and color. The size of the particles was estimated by fitting an ellipse and then using the major and minor axis.…”

“…All the previous methods based on image analysis have required dedicated equipment as special staining dyes and light (Prata et al, 2019), flat scanner (Mukhanov et al, 2019) (Gauci et al, 2019) (Lorenzo-Navarro et al, 2020 or microscopy images (Wegmayr et al, 2020). The methodology proposed in this work is based on images taken with cameras or mobile phones with a resolution of at least 16 million pixels.…”