Microplastic debris in major water systems has become a significant issue in recent years. Monitoring microplastic pollution and evaluating its health risks are largescale jobs. Bin Shi, a PhD student from the Department of Material Science & Engineering, strives to reduce these challenges through his project, Automatic quantification and classification of microplastics in scanning electron micrographs via deep learning.
Shi’s project, supervised by Professors Jane Howe (ChemE, MSE), Elodie Passeport (ChemE, CivMin) and Dwayne Miller (Chem, Physics), facilities microplastic quantification and classification with high accuracy, including instances when the microplastics are densely packed or imaged in complex environments with poor experimental settings.
“We have collected microplastics in various shapes and chemical compositions from daily supplies, such as washing and dryer machines, packing film, face and body wash, masks, cookware, and lunchboxes just to name a few examples. We have imaged these supplies by scanning electron microscopy (SEM). This offers greater depth of microplastics at a wider range of magnification than visible-light microscopy or a digital camera. Also, SEM has the potential to go down to the scale of nanoplastics. The smaller microplastics we can observe, the more microplastics we can take into account,” says Shi.
This approach has allowed Shi and his team to create the FIRST labelled open-source SEM dataset of microplastics for image segmentation. Additionally, Shi and his team applied deep-learning methods to automate and facilitate quantification and classification of microplastics in SEM images, which achieved a significantly improved performance than conventional methods.
“With the new data from our project, we have a way to further simplify microplastic quantification and classification. This helps us address the growing challenges that microplastics have posed to water systems, which affect human health, economic development, and political stability,” explains Shi.
Details of Shi’s project are available on the ScienceDirect website and will soon be published in the Science of the Total Environment Journal. Shi’s project was made possible through contributions from the Institute for Water Innovation Waterseed Program, Ontario Centre for the Characterisation of Advanced Materials research space, as well as students and professors from U of T Engineering. “Alone we can do so little, together we can do so much,” says Shi.