Skip to Content
Find More Like This
Return to Search

Automated Image Analysis of Fibers

Automatic Nanofiber Characterization and Recognition Software

Argonne National Laboratory

Contact ANL About This Technology

Image with recognized fiber edges<br />
<br />
Diameter - Measure between each yellow and red tail.
Image with recognized fiber edges

Diameter - Measure between each yellow and red tail.

Error = 4.7%
Error = 4.7%

Technology Marketing Summary

The present invention comprises a method of automatic characterization of electrospun nanofiber assemblies. The novel system uses existing edge detection algorithms with a Hough transform to recognize and measure each fiber orientation angle and diameter for assembly characterization.


This new system is designed specifically to have an intuitive approach of recognizing each fiber for robust performance on nanofiber assemblies with measurement difficulties such as beads, curved fibers, and varying area density. Mean diameter performance is excellent for clean images (<5% deviation from manual measurement), decent for various challenging issues (<20% deviation), and relatively poor for blurry images (<50% deviation). Fiber alignment and diameter standard deviation also show good agreement with manual results. In most of the cases our method outperforms methods described in open literatures, especially for images with challenging but common and realistic issues, which enabled our method to be practically useful in real life research as a dependable tool.

•Fundamentally different approach
•Intuitive algorithm – a human like approach yields more reasonable measurements: no weighing error, better noise rejection, no issue with cross sections etc.
•Higher performance for all types of images
•Works well on majority of real life images (~5% error for ideal, ~20% error for others), an enabler for real world applications
•Alignment data obtained with no additional computational cost
•An convenient semi-automatic mode with visual verification/modification on-the-flight that could significantly improve performance in particular challenging images, thanks to the human-like approach
•Very computational efficient – fast response suitable for in-situ quality control (10-100 time faster than existing methods)
Applications and Industries
  • Advanced Manufacturing
  • Nanomaterials
  • Nanocomposites
  • Nanomaterial Research
More Information

Currently being evaluated by commercial entitities.

Technology Status
Technology IDDevelopment StageAvailabilityPublishedLast Updated
SF-14-107Production - Currently be Used in Analyzing Fiber Matrixes at the LabAvailable03/30/201508/21/2015

Contact ANL About This Technology

To: Elizabeth Jordan<>