高电压技术2026，Vol.52Issue(7)：2973-2985,13.DOI:10.13336/j.1003-6520.hve.20260604
领域知识引导的视觉语言模型配电线路金具锈蚀检测方法
Domain Knowledge-guided Vision-language Model for Corrosion Detection of Distribution Line Fittings
摘要
Abstract
To address the issues such as difficulty in defect representation,insufficient semantic information,and am-biguous class boundaries in corrosion detection of distribution-line fittings,a domain knowledge-guided vision-language detection method is proposed.First,a low-contrast feature enhancement module is designed to enhance textures,bounda-ries,and key channel responses of low-contrast corrosion regions through multi-scale directional extraction,boundary-guided modulation,and channel adaptive recalibration.Second,a knowledge-guided representation module is constructed to transform distribution-line domain knowledge into textual prompts and to dynamically optimize text em-beddings with a vision-conditioned prompt adapter,improving the model's semantic understanding of corrosion patterns and fitting states.Finally,a cross-modal semantic alignment strategy is introduced to constrain visual-semantic consisten-cy through region-text contrastive classification loss and semantic alignment regularization,enhancing discrimination of similar categories.The experimental results show that the proposed method achieves 87.1％mean average precision(mAP50)on a self-built defect dataset of distribution-line fittings.In open-semantic detection and zero-shot generalization experiments,the mAP50 values reach 81.2％and 44.2％,respectively,with an inference speed of 34.9 frames per second.The proposed method can achieve a balance among detection accuracy,open-category generalization and inference effi-ciency,providing a technical support for defect detection of distribution-line fittings.关键词
配电线路/缺陷检测/视觉语言模型/领域知识/低对比特征增强/跨模态对齐/开放词汇检测Key words
distribution lines/defect detection/vision-language model/domain knowledge/low-contrast feature en-hancement/cross-modal alignment/open-vocabulary detection引用本文复制引用
赵振兵,田本喜,高凯鹏,唐辰康,李浩鹏..领域知识引导的视觉语言模型配电线路金具锈蚀检测方法[J].高电压技术,2026,52(7):2973-2985,13.基金项目
国家自然科学基金(62571189 （）
62373151 （）
62371188 （）
62303184) （）
中央高校基本科研业务费专项资金(2023JC006 （）
2025MS118).Project supported by National Natural Science Foundation of China(62571189,62373151,62371188,62303184),Fundamental Research Funds for the Central Universities(2023JC006,2025MS118). （62571189,62373151,62371188,62303184）
