We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Designing a Low-Cost Microcontroller-Based Rover for Microplastic Detection Using Deep-Learning Image Detection and Near-Infrared Spectroscopy
Summary
Researchers designed a low-cost microcontroller-based rover for detecting nurdle microplastics in shoreline environments, integrating a compressed deep-learning object detection model trained on 150 images of polyethylene pellets with an AS7263 near-infrared sensor for spectral confirmation of polyethylene. The Raspberry Pi 3-based system demonstrated efficient microplastic identification across varying lighting conditions and burial depths in sand.
This research article presents a potential solution to microplastic accumulation, with a specific focus on nurdles, which are small-rounded plastics used extensively in manufacturing processes. These nurdles have emerged as a major contributor to the buildup of smaller microplastics in our environment. To combat this problem, this study presents an innovative solution in the form of a low-cost rover designed to track and detect these particles in shoreline areas. The research methodology employed in this study encompasses several key components. First, a comprehensive dataset consisting of 150 images depicting polyethylene plastic pellets buried under sand, varying in lighting conditions and quantities, is used to train a compressed object detection model. This model, when executed on the Raspberry Pi 3, enables efficient and accurate identification of microplastic particles. To further enhance the detection capabilities, an AS7263 NIR sensor is used in conjunction with the Arduino BLE Sense 33. This integration allows for the measurement of near-infrared reflectance in polyethylene, thereby serving as an additional metric for microplastic detection. By analyzing the spectral characteristics of polyethylene, the system gains further insights into the presence and concentration of microplastics. This proposed innovation presents a promising solution for combating the adverse effects of microplastic pollution. The development of a low-cost rover, coupled with deep-learning image detection and near-infrared spectroscopy, offers a powerful and cost-effective approach to identifying and monitoring microplastic contamination in shoreline environments. This research contributes to the ongoing efforts to mitigate the environmental impact of microplastics and paves the way for future advancements in this field.
Sign in to start a discussion.
More Papers Like This
Comparison of learning models to predict LDPE, PET, and ABS concentrations in beach sediment based on spectral reflectance
Researchers compared machine learning models to predict concentrations of LDPE, PET, and ABS microplastics in beach sediments using visible-near-infrared spectral reflectance, demonstrating that spectroscopic methods can efficiently estimate microplastic pollution in understudied terrestrial and coastal environments.
Identification of Polymers with a Small Data Set of Mid-infrared Spectra: A Comparison between Machine Learning and Deep Learning Models
Researchers compared multiple machine learning and deep learning models for identifying polymer types from mid-infrared spectral data using a small reference dataset, finding that certain deep learning architectures outperformed traditional methods even with limited training examples, supporting automated microplastic identification.
Quantifying Marine Plastic Debris in a Beach Environment Using Spectral Analysis
Researchers analyzed shortwave infrared reflectance spectra of weathered marine plastic debris on sandy beaches, finding that polymer type significantly influences detection capability at sub-pixel surface covers relevant to remote sensing applications.
Polymer Type Identification of Marine Plastic Litter Using a Miniature Near-Infrared Spectrometer (MicroNIR)
Researchers tested a miniature near-infrared spectrometer (MicroNIR) for rapidly identifying polymer types in marine plastic litter collected from beaches, finding it could accurately distinguish common plastics like polyethylene and polypropylene. Low-cost, portable identification tools are important for large-scale monitoring of marine plastic pollution.
Unmanned Vehicle and Hyperspectral Imager for a More Rapid Microplastics Sampling and Analysis
Researchers tested a combination of an autonomous surface vehicle and a near-infrared hyperspectral imager to rapidly sample and identify microplastics on the Norwegian coast. Results compared favorably with standard FTIR analysis and demonstrated a repeatable method for assessing spatially variable microplastic concentrations in the marine environment.