Automatic ply-specific analyses of CFRP micrographs using shortest-path-based ply distinction

2026-06-17 · Source: Computer Vision and Pattern Recognition · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Materials Science & Engineering · Depth: Expert, quick

Summary

An automated method has been developed to distinguish ply instances within semantic segmentation masks of high-resolution carbon-fiber reinforced polymer (CFRP) micrographs. This approach interprets the segmentation mask as a graph, where pixels are vertices, and then employs a shortest-path algorithm to identify ply-separating paths. This technique effectively bridges the gap between semantic segmentation and ply instance segmentation by leveraging global information. The method has been successfully applied to high-resolution micrographs exhibiting diverse characteristics, including artificially added gaps in single or multiple plies, varying stacking sequences, and ply-traversing cracks. By assigning each fiber pixel to a specific ply based on these calculated paths, the system enables comprehensive, quantitative ply analysis. This analysis covers microstructural properties such as local fiber volume fraction, as well as locally resolved ply and interleaf layer thickness, providing insights into manufacturing-induced inhomogeneities and their link to mechanical properties.

Key takeaway

For materials scientists or quality control engineers analyzing CFRP composites, this automated ply distinction method offers a critical tool to enhance microstructural analysis. You can precisely quantify local fiber volume fraction and ply thickness, directly linking these to manufacturing parameters and potential mechanical property impacts. This allows you to identify process-induced inhomogeneities earlier, improving quality assurance and informing material design decisions.

Key insights

A shortest-path algorithm on graph-represented segmentation masks enables automated ply instance distinction in CFRP micrographs.

Principles

• Graph interpretation of segmentation masks enables global information use.
• Ply instance segmentation can reveal manufacturing inhomogeneities.

Method

Interpret semantic segmentation masks as graphs with pixels as vertices, then apply a shortest-path algorithm to find ply-separating paths for instance distinction and subsequent quantitative analysis.

In practice

• Quantify local fiber volume fraction in CFRP plies.
• Measure local ply and interleaf layer thickness.
• Identify manufacturing-induced microstructural imperfections.

Topics

• Carbon-Fiber Reinforced Polymer
• Micrograph Analysis
• Semantic Segmentation
• Instance Segmentation
• Shortest Path Algorithms
• Quality Control

Best for: AI Scientist, Research Scientist, Computer Vision Engineer

Related on AIssential

• STEP-Parts: Geometric Partitioning of Boundary Representations for Large-Scale CAD Processing — STEP-Parts extracts stable geometric instance partitions from B-Reps, preserving analytic structure for CAD learning.
• $${\bf{Micro}}{{\mathbb{S}}}{\bf{plit}}$$ Micro S plit : semantic unmixing of fluorescent microscopy data — MicroSplit enables simultaneous multi-structure imaging in one channel, providing denoised, unmixed outputs with uncertainty quantification.
• Towards Effective Waste Segmentation for Automated Waste Recycling in Cluttered Background — The network combines spatial and spectral domains with an AFEM for robust waste segmentation in cluttered AWR environments.
• VoxelDiffusionCut: Non-destructive Internal-part Extraction via Iterative Cutting and Structure Estimation — Iterative internal structure estimation via diffusion models enables non-destructive part extraction from unknown products.
• Topological texture analysis of microscopy images of dynamic casein gelation and its relation to rheological properties — The toolbox combines TDA, DBC, MFP, and LBP to quantify dynamic microstructural changes in gelation, correlating with rheological properties.
• Local graph estimation with pathwise false discovery control — Local graph estimation with pathwise feature selection (PFS) infers interpretable substructures around target variables.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computer Vision and Pattern Recognition.