[Poster Presentation]Mask R-CNN-based bushing segmentation with refined feature extraction and precise localization

Mask R-CNN-based bushing segmentation with refined feature extraction and precise localization
ID:11 Submission ID:12 View Protection:ATTENDEE Updated Time:2025-10-11 21:50:44 Hits:207 Poster Presentation

Start Time:2025-11-09 09:00 (Asia/Shanghai)

Duration:1min

Session:[P] Poster presentation » [P6] 6.AI-driven technology

No files

Abstract
电力设备的准确定位和分割对于自动化检测系统至关重要,特别是在检测结构和热异常方面。然而,由于设备组件之间的复杂背景和高度的类间相似性,这项任务仍然具有挑战性。本研究提出了一个增强的掩码 R-CNN 框架,通过两个关键改进来应对这些挑战:(1) 集成软非极大抑制 (Soft-NMS) 以细化重叠对象的边界框回归,(2) 优化特征提取以增强判别能力。在套管数据集上的实验结果表明,与基线掩模R-CNN模型相比,所提方法在IoU(交集交集)阈值为0.5时,平均精度提高了6%。该框架在各种成像条件下(包括红外和可见光谱)表现出强大的性能,同时保持计算效率。通过解决现有方法中的关键局限性,这项工作为电力设备细分提供了实用的解决方案,并为电气工业应用中的可扩展缺陷检测奠定了基础。
Keywords
improved MASK R-CNN,power equipment segmentation,Transformer bushing
Speaker
Yuhan You
Student Huazhong university of science and technology

Submission Author
Yuhan You Huazhong university of science and technology
Yong Yang Huazhong University of Science and Technology
Chuan Li Huazhong University of Science and Technology
Comment submit
Verification code Change another
All comments

Contact us

CIYCEE 2025 Official E-mail:

ciycee2025@163.com

 

WeChat public account: 

IEEE IAS SWJTU Student Branch