Attention-Guided Autoencoder Fusion for Insulator Defect Detection Using UAV Transmission-Line Imaging

2026-06-03 · Source: Computer Vision and Pattern Recognition · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Expert, quick

Summary

AE-YOLO, an Attention-Guided AutoEncoder-Enhanced YOLO framework, addresses challenges in automated defect detection for high-voltage transmission-line insulators using Unmanned Aerial Vehicle (UAV) imagery, specifically severe class imbalance, large scale variation, and small defect instances. The architecture integrates lightweight bottleneck autoencoders within a Feature Pyramid Network-Path Aggregation Network (FPN-PAN) neck to preserve anomaly-sensitive information during multi-scale feature fusion. Convolutional Block Attention Modules (CBAM) enhance feature discrimination, while a variance-maximizing autoencoder regularization strategy encourages diverse, defect-discriminative latent representations. The network trains with a unified objective combining focal loss, Complete IoU (CIoU) loss, and autoencoder regularization. During inference, Weighted Boxes Fusion (WBF) combines predictions from YOLOv8, YOLOv10, and YOLO11, augmented by an autoencoder-guided confidence boosting mechanism for rare defect categories. Experiments on the Insulator-Defect Detection dataset show AE-YOLO with an EfficientNetV2 backbone achieves 95.10% mAP at 0.5, 96.40% precision, and 93.80% recall, surpassing the strongest YOLO-family baseline by 5.0 points in mAP at 0.5 and 6.7 points in recall.

Key takeaway

For Computer Vision Engineers developing robust defect detection systems for critical infrastructure like transmission lines, AE-YOLO presents a highly effective framework. You should consider integrating lightweight bottleneck autoencoders within your FPN-PAN neck and Convolutional Block Attention Modules throughout your backbone to improve anomaly sensitivity and feature discrimination. Adopting multi-model fusion with confidence boosting, as demonstrated by AE-YOLO's 95.10% mAP at 0.5, can significantly enhance detection accuracy and recall for rare defect categories in UAV imagery.

Key insights

AE-YOLO integrates autoencoders and attention mechanisms into YOLO for robust UAV-based insulator defect detection, improving mAP and recall.

Principles

• Autoencoders preserve anomaly-sensitive information.
• Attention modules enhance feature discrimination.
• Multi-model fusion improves detection robustness.

Method

AE-YOLO integrates bottleneck autoencoders into FPN-PAN and CBAMs into the backbone. It uses variance-maximizing regularization and a unified loss (focal, CIoU, autoencoder regularization). Inference combines YOLOv8, YOLOv10, YOLO11 via WBF with autoencoder-guided confidence boosting.

In practice

• Inspect transmission lines via UAVs.
• Monitor insulator defects effectively.
• Enhance rare defect category detection.

Topics

• Insulator Defect Detection
• UAV Imaging
• Attention Mechanisms
• Autoencoders
• YOLO Frameworks
• Multi-model Fusion

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

Related on AIssential

• Attention-Guided Autoencoder Fusion for Insulator Defect Detection Using UAV Transmission-Line Imaging — Integrating autoencoders and attention into YOLO enhances UAV-based insulator defect detection, especially for small, rare anomalies.
• Breaking the Bottleneck: Achieving Native NMS-Free Inference with YOLO26 — This code provides a full pipeline for ONNX-based YOLOv8n object detection from image loading to result saving.
• TAP into the Patch Tokens: Leveraging Vision Foundation Model Features for AI-Generated Image Detection — Modern vision foundation models significantly outperform CLIP for AI-generated image detection.
• Rethinking Infrastructure Inspection as Image Difference Classification: A Traffic Sign Case Study — Image difference classification with instruction-based models reduces data dependency for infrastructure defect detection.
• HYolo: An Intelligent IoT-Based Object Detection System Using Hypergraph Learning — HYolo integrates hypergraph learning into YOLO to model complex high-order relationships, improving object detection in IoT.
• Masked Autoencoder for Image Reconstruction using Vision Transformers (PyTorch) — Masked Autoencoders with Vision Transformers learn robust image representations by reconstructing heavily masked image patches.

Open in AIssential →

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