What does this research mean for the field?

Integrating UAV imaging with the YOLOv8 deep learning model for automated visual inspection of power line infrastructure achieves 88.5% mAP@0.5 and real-time inference, enabling a threefold increase in inspection coverage and up to a 70% reduction in defect remediation time. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

The research aims to develop a deep learning framework for efficient visual inspection of overhead power lines using UAVs.

March 18, 2026Open Access

Deep Learning-Based Visual Analytics for Efficiency and Safety Optimization in Power Infrastructure

Key Points

The research aims to develop a deep learning framework for efficient visual inspection of overhead power lines using UAVs.
Developed a framework using UAVs for visual inspection.
Integrated YOLOv8, EfficientDet-D2, and Faster R-CNN for defect detection.
Utilized datasets InsPLAD, TTPLA, and MPID for training and validation.
Conducted experiments to evaluate performance under various conditions.
YOLOv8 achieved 88.5% mAP@0.5 accuracy with real-time inference over 50 FPS on GPU.
Increased inspection coverage capacity by threefold.
Reduced defect remediation time by up to 70%.

Abstract

The paper presents a comprehensive deep learning-based framework for automated visual inspection of overhead power line infrastructure using unmanned aerial vehicles. Traditional manual and helicopter inspections are costly, time-consuming, and hazardous for maintenance personnel. The proposed approach integrates UAV imaging with advanced computer vision models such as YOLOv8, EfficientDet-D2, and Faster R-CNN to automatically detect defects in critical components, including insulators, conductors, and transmission towers. Several open datasets (InsPLAD, TTPLA, MPID) were used for training and validation, ensuring robustness under diverse lighting and environmental conditions. Experimental results demonstrate that YOLOv8 achieved the best performance, reaching 88.5% mAP@0.5 with real-time inference capabilities (over 50 FPS on GPU). The system significantly enhances inspection efficiency, allowing for a threefold increase in coverage capacity and an up to 70% reduction in defect remediation time. The integration of AI-powered visual analytics with maintenance and SCADA systems enables a shift from reactive to predictive maintenance, improving the safety, reliability, and resilience of power transmission infrastructure.

Read Full Paperexternally

Mark Helpful

Bookmark

Relay

View Full Paper