Detection and classification of failed fiber reinforced polymer composite samples under mechanical loading using YOLO11 on microscopic SEM images

The fracture surfaces of fiber-reinforced polymer (FRP) composites have multiple scale failure features that are hard to interpret manually by scanning electron microscope (SEM) and are slow to conduct. The fracture surfaces of FRP composites have multiple scale failure features that are hard to interpret manually by scanning electron microscope (SEM) and slow to conduct. This proof-of-concept study compares an anchor-free single-stage detector, YOLO11n, to localize and classify six fractographic features: fiber breakage, fiber pull-out, fiber–matrix debonding, delamination, voids and matrix debris, in SEM images of carbon- and glass-fiber/epoxy specimens failed in tension, three-point bending and short-beam shear. The data consists of 125 images with 37 specimens, with 1318 labelled boxes following written labelling criteria. The weights for the model were initialized from COCO and were updated using the Adam optimizer. YOLO11n was able to obtain [email protected] = 0.821 with 6.5 ms per image on a preliminary image-level split, but the number of images per subset is not consistent with 125 images and the results may be optimistic due to the fact that images from one specimen can be split into multiple subsets. Multiple seeds, a controlled comparison between it and YOLOv5n, YOLOv8n and YOLOv9t, per-class ablations and an error analysis were performed at the level of the specimen and the groups. The study shows that it is possible to apply the method of detector-assisted laboratory fractography. It does not provide any structural-health-monitoring or safety-critical capability, and is not tested on other material systems or imaging conditions.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2026-11-15
DOI
https://doi.org/10.1038/s41598-026-78824-4
Primary Topic
Industrial Vision Systems and Defect Detection
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Detection and classification of failed fiber reinforced polymer composite samples under mechanical loading using YOLO11 on microscopic SEM images

T. M. Yunus Khan, Wahaj Ahmad Khan, B. Sandeep, Yosef Jazaa et al.
Scientific Reports
Industrial Vision Systems and Defect Detection
article

Detection and classification of failed fiber reinforced polymer composite samples under mechanical loading using YOLO11 on microscopic SEM images

T. M. Yunus Khan, Wahaj Ahmad Khan, B. Sandeep, Yosef Jazaa, Shajahan Aboobacker, B. Roopa, J. Pallavi, A. Vamshi
article en

Abstract

The fracture surfaces of fiber-reinforced polymer (FRP) composites have multiple scale failure features that are hard to interpret manually by scanning electron microscope (SEM) and are slow to conduct. The fracture surfaces of FRP composites have multiple scale failure features that are hard to interpret manually by scanning electron microscope (SEM) and slow to conduct. This proof-of-concept study compares an anchor-free single-stage detector, YOLO11n, to localize and classify six fractographic features: fiber breakage, fiber pull-out, fiber–matrix debonding, delamination, voids and matrix debris, in SEM images of carbon- and glass-fiber/epoxy specimens failed in tension, three-point bending and short-beam shear. The data consists of 125 images with 37 specimens, with 1318 labelled boxes following written labelling criteria. The weights for the model were initialized from COCO and were updated using the Adam optimizer. YOLO11n was able to obtain [email protected] = 0.821 with 6.5 ms per image on a preliminary image-level split, but the number of images per subset is not consistent with 125 images and the results may be optimistic due to the fact that images from one specimen can be split into multiple subsets. Multiple seeds, a controlled comparison between it and YOLOv5n, YOLOv8n and YOLOv9t, per-class ablations and an error analysis were performed at the level of the specimen and the groups. The study shows that it is possible to apply the method of detector-assisted laboratory fractography. It does not provide any structural-health-monitoring or safety-critical capability, and is not tested on other material systems or imaging conditions.

Scientific Reports
National Institute of Technology Karnataka (IN), Dire Dawa University (ET), King Khalid University (SA)
Industrial Vision Systems and Defect Detection
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon