Related Experiment Video
Updated: Aug 7, 2026
09:27
Measuring the Shape and Size of Activated Sludge Particles Immobilized in Agar with an Open Source Software Pipeline
Published on: January 30, 2019
Automated SEM-based nanoparticle metrology for materials characterization via segmentation and robust scale-bar
Xueyi Huang1, Yuzhong Yin2, Jinghao Hu2
1School of Computer Science and Information Engineering, Hefei University of Technology Hefei 230009 China jianliu@hfut.edu.cn.
RSC Advances
|August 6, 2026
Summary
We developed an automated workflow for nanoparticle metrology using scanning electron microscopy (SEM). This method accurately analyzes particle size and shape, improving materials research reproducibility.
Area of Science:
- Materials Science
- Nanotechnology
- Image Analysis
Background:
- Quantitative nanoparticle metrology is crucial for materials science applications.
- Current scanning electron microscopy (SEM) methods face challenges in high-throughput analysis due to ambiguous boundaries, agglomeration, and scale calibration issues.
Purpose of the Study:
- To develop an automated SEM-based nanoparticle metrology workflow.
- To enable accurate, reproducible, and high-throughput characterization of nanoparticle size and morphology.
Main Methods:
- Integration of Multi-scale U-shaped Kolmogorov-Arnold Network (MU-KAN) for image segmentation.
- Utilized YOLOv11 for scale-bar localization and OCR for scale annotation parsing.
- Developed a workflow to convert raw SEM images into scale-calibrated particle statistics.
Main Results:
- MU-KAN achieved high performance on benchmark datasets (IoU: 0.9296, 0.8243; F1-score: 0.9630, 0.9029).
- The scale recovery module demonstrated a low mean relative error of 3.8604%.
- The workflow provides comprehensive particle descriptors like size, circularity, and shape statistics.
Conclusions:
- The automated workflow significantly enhances nanoparticle characterization accuracy and reproducibility.
- This method facilitates structure-property relationship studies in diverse materials research fields.
- Enables automated, physically meaningful analysis of nanoparticles from SEM images.
