0
Article ? AI-assigned paper type based on the abstract. Classification may not be perfect — flag errors using the feedback button. Tier 2 ? Original research — experimental, observational, or case-control study. Direct primary evidence. Marine & Wildlife Policy & Risk Sign in to save

Development of Microplastics Detector and Quantifier Utilizing Deep Learning Based Algorithm

2024
Mark Angelo C. Regoso, Melissa B. Galang, John E. Mendoza, Jhoana C. Suelo, Sem B. Jalotjot, Herbert V. Villaruel

Summary

Researchers developed a microplastics detector and quantifier using deep learning-based image analysis, training a neural network to identify and count microplastic particles in microscopic images. The system achieved high accuracy and offers a faster, more objective alternative to manual counting.

Microplastics are tiny plastic particles with a size smaller than 5mm. These are considered ubiquitous environmental contaminants as they pollute shorelines and bodies of water globally. Also, there are potentially hazardous effects of alternate ingestion to humans and marine life. In alignment with SDG 14: Life Below Water, the researchers consider the Barangay Binuangan to correspond best in this endeavor. Binuangan is a barangay in the Philippine province of Bulacan's coastal municipality of Obando. The proposed solution is to create a prototype that utilizes a deep-learning-based algorithm that detects and quantifies microplastics. The system uses a deep-learning model; You Only Look Once version 4 (YOLOv4). The researchers used a manta trawl to obtain the residue for specific project design requirements, which is crucial as it leverages the accuracy percentage in detecting the microplastics. The target average precision (AP) and mean average precision (mAP) are greater than 85% and 65%, respectively. Using YOLOv4 achieved an AP for microplastics class of 92.25% and mAP of 70.52%, a 7.83% and 6.21 % increase to AP and mAP, respectively than that of using YOLOv4-tiny on a real-time detection using a microscopic camera.

Share this paper