Intelligent detection of transmission line icing in extreme conditions: a Multi-Dimensional Attention-enhanced DETR approach

Résumé fourni par la source

Abstract A novel detection algorithm leveraging the Multi-Dimensional Attention-enhanced DETR (MDAL-DETR) framework is introduced to identify ice accumulation on transmission lines in challenging and extreme conditions. The proposed Thin-head feature extraction network, incorporating a lightweight model, effectively retains spatial information while progressively propagating hidden dependencies across channels. To address the imbalance between positive and negative samples, a loss function (Shape-IOU) merging attention to morphological cross-ratio and normalization is proposed. Experimental results validate that the proposed network accurately detects ice-covered sections of power transmission lines in harsh environments, minimizing both false alarms and missed detections. The approach achieves a mean Average Precision (mAP) of 97.85%, showcasing its potential for real-time monitoring and early detection of icing hazards on transmission lines.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.