Insulators are vital components of high-voltage power transmission systems, where undetected defects can lead to catastrophic failures and significant economic losses. Accurate and timely detection of insulator defects (IDs) under diverse environmental conditions is critical for ensuring system reliability. This study presents Transformer-Enhanced YOLOv8 (TE-YOLOv8), a novel hybrid deep learning framework designed to address the challenges of detecting small, complex defects in transmission line inspections. TE-YOLOv8 integrates transformer-based attention mechanisms with the advanced YOLOv8 architecture, introducing several key innovations that enhance its performance. Specifically, it incorporates Global Convolution (GConv) modules to capture extended spatial context for improved feature extraction, C3f-Global Pooling Fusion (C3f-GPF) modules to amplify discriminative features, and Multiscale Information Fusion (MSIF) modules with learnable weights for adaptive multi-scale detection. Additionally, it utilizes Weighted Feature Information Fusion (WFIF) modules for channel-wise attention to refine feature representation, and a Transformer-enhanced neck architecture to model global dependencies and provide enhanced contextual understanding. To improve localization precision and accelerate convergence, the framework adopts the SCYLLA-IoU (SIoU) loss function. Extensive experimental validation on the IDID and CPLID datasets demonstrates that TE-YOLOv8 achieves mean average precision (mAP) scores of 94.2% and 93.8%, respectively, representing improvements of 4.9% and 5.1% over the baseline YOLOv8, and 1.9% and 2.0% over TE-YOLOV8, while maintaining real-time inference at 82 frames per second. Ablation studies, precision-recall curves, and visualization analyses further confirm the effectiveness of TE-YOLOv8 in detecting insulator defects under challenging operational conditions.
Farooq et al. (2026) studied this question.