What question did this study set out to answer?

To develop and evaluate a hybrid CNN-GRU framework that automates surface defect classification and data-driven severity grading for wind turbine blades using drone inspection imagery.

August 19, 2026Discover Applied SciencesOpen Access

Hybrid CNN-GRU framework for wind turbine blade defect classification and data-driven severity assessment for predictive maintenance

View Full Paper

Ask AI

Bookmark

Authors

AKAshwitha K SSSurendra Shetty SVSrinivas Byatarayanapura Venkataswamy

Discussion

Member takes

Overview

Experimental study demonstrates accurate turbine blade defect detection from drone imagery using a hybrid CNN-GRU, highlighting practical predictive maintenance.

Key Points

To develop and evaluate a hybrid CNN-GRU framework that automates surface defect classification and data-driven severity grading for wind turbine blades using drone inspection imagery.
Trained a hybrid model integrating MobileNetV2 for spatial feature extraction with a Gated Recurrent Unit (GRU) on a dataset of 2,995 drone-captured wind turbine blade inspection images.
Determined crack severity levels directly from the statistical distribution of bounding box annotation areas rather than subjective thresholds.
Compared performance against five baseline models (Coordinate Attention CNN, MobileNetV2, ResNet50, EfficientNetB0, and YOLOv5s) and conducted ablation testing.
The CNN-GRU framework achieved 95.67% accuracy, an F1-Score of 87.13%, and an area under the receiver operating characteristic curve (AUROC) of 0.9756 on the held-out test set.
Ablation of the GRU component decreased the F1-Score from 87.13% to 47.50% and reduced dirt-class sensitivity from 89.80% to 38.78%.

Cite This Study

K et al. (2026) studied this question.

synapsesocial.com/papers/6a85638803308d306e2d6bfb https://doi.org/10.1007/s42452-026-09391-6

View Full Paper

Ask AI

Bookmark

August 19, 2026Discover Applied SciencesOpen Access

Hybrid CNN-GRU framework for wind turbine blade defect classification and data-driven severity assessment for predictive maintenance

View Full Paper

Ask AI

Bookmark

Authors

AKAshwitha K SSSurendra Shetty SVSrinivas Byatarayanapura Venkataswamy

Discussion

Member takes

Overview

Experimental study demonstrates accurate turbine blade defect detection from drone imagery using a hybrid CNN-GRU, highlighting practical predictive maintenance.

Key Points

To develop and evaluate a hybrid CNN-GRU framework that automates surface defect classification and data-driven severity grading for wind turbine blades using drone inspection imagery.
Trained a hybrid model integrating MobileNetV2 for spatial feature extraction with a Gated Recurrent Unit (GRU) on a dataset of 2,995 drone-captured wind turbine blade inspection images.
Determined crack severity levels directly from the statistical distribution of bounding box annotation areas rather than subjective thresholds.
Compared performance against five baseline models (Coordinate Attention CNN, MobileNetV2, ResNet50, EfficientNetB0, and YOLOv5s) and conducted ablation testing.
The CNN-GRU framework achieved 95.67% accuracy, an F1-Score of 87.13%, and an area under the receiver operating characteristic curve (AUROC) of 0.9756 on the held-out test set.
Ablation of the GRU component decreased the F1-Score from 87.13% to 47.50% and reduced dirt-class sensitivity from 89.80% to 38.78%.

Cite This Study

K et al. (2026) studied this question.

synapsesocial.com/papers/6a85638803308d306e2d6bfb https://doi.org/10.1007/s42452-026-09391-6

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark