高校学生人工智能安全素养现状调查与提升路径研究
摘要
关键词
全文:
PDF参考
乔雪峰.从工具赋能到智能协同:生成式人工智能驱动的教育模式转型[J].南京社会科学,2025,(01):126-134.DOI:10.15937/j.cnki.issn1001-8263.2025.01.013.
孙立会,周亮.生成式人工智能赋能教育变革的逻辑——基于新质生产力的视角[J].教育研究,2024,45(10):38-49.
辛征,郝丽丽,侯传晶,等.AI时代大学生国家安全素养培养策略研究[J].电气电子教学学报,2025,47(01):87-90.
郑腾.人工智能赋能高校治理的背景阐释、潜在风险与改善路径[J].黑龙江高教研究,2025,43(09):1-7.DOI:10.19903/j.cnki.CN23-1074/G.2025.09.001.Design of an Intelligent Safety Monitoring System for Discharge Glows in Power Equipment Based on Improved YOLOv11nJiaxin Yue Yufeng Wang Benqing Ma Yao Li Xinyi HuSchool of Electronics and Information Engineering, Liaoning University of Science and Technology, Anshan, Liaoning, 114051, China AbstractTo address the core pain points in current discharge monitoring of power system insulators—namely low inspection efficiency, poor safety, and high missed detection rates—this paper designs an intelligent safety monitoring system for power equipment discharge based on an improved YOLOv11n. By introducing the CBAM attention mechanism into the YOLOv11n architecture, optimizing anchor box ratios, and integrating lightweight convolutional modules, th
e model enhances its ability to extract micro-discharge features and adapt to diverse scenarios, enabling data extraction, type classification, and severity assessment of discharge areas. The system establishes a full-cycle development framework encompassing “data collection, model training, evaluation optimization, and hardware deployment,” suitable for various power working conditions such as humid heat, high temperature, and normal environments. Testing results demonstrate that the system reduces missed detection rates for the three types of discharge by over 80% compared to manual inspections, with a single-image detection latency of ≤120ms on an ordinary i5 processor terminal. It effectively supports preventive maintenance of power equipment, providing intelligen
t safety assurance for stable power system operation.KeywordsImproved YOLOv11n; Power Equipment; Discharge Glow Monitoring; Intelligent Safety; CBAM Attention Mechanism基于改进 YOLOv11n 的电力设备放电微光智能安全监测系统设计岳佳欣王玉峰马本卿李垚胡馨仪辽宁科技大学电子与信息工程学院,中国·辽宁鞍山 114051摘要针对当前电力系统绝缘子缺陷放电、高压弧光、电弧老化局部放电三类放电监测中存在的巡检效率低、安全性差、漏检率高的核心痛点,本文设计了一套基于改进YOLOv11n的电力设备放电微光智能安全监测系统。通过在YOLOv11n基础架构中引入CBAM注意力机制、优化锚框比例、搭载轻量化卷积模块,提升模型对微光放电特征的提取能力与多场景适配性,实现放电区域数据提取、类型区分与等级判定。系统构建了“数据采集-模型训练-评估优化-硬件部署”全流程开发体系,适配湿热、高温、常规等多电力工况,经测试验证,三类放电检测漏检率较人工巡检降低80%以上,普通i5处理器终端单图检测时延≤120ms,可有效支撑电力设备预防性维护,为电力系统稳定运行提供智能化安全保障。关键词改进YOLOv11n;电力设备;放电微光监测;智能安全;CBAM注意力机制【基金项目】辽宁科技大学大学生创新创业训练计划项目基金资助。【作者简介】岳佳欣(2005—),女,本科,中国河南许昌人。1 引言电力设备放电(如绝缘子缺陷放电、高压弧光等)是引发电力系统跳闸、设备损坏甚至大面积停电的核心隐患,其监测工作直接关系到电力系统的安全稳定运行。当前电力行业主流的放电监测方式以人工巡检为主,依赖运维人员现场排查、肉眼识别,存在明显局限性。随着深度学习技术在目标检测领域的快速发展,YOLO系列算法凭借高效的实时检测能力与轻量化架构,已在工业监测、安防监控等领域得到广泛应用。YOLOv11n作为该系列最新的轻量化模型,具备运算效率高、部署成本低的优势,但其原始架构针对电力设备放电微光场景的适配性不足,存在微弱特征捕捉不精准、多目标(小目标绝缘子放电、中大型目标高压弧光)检测适配性差等问题。
DOI: http://dx.doi.org/10.12345/bdai.v7i2.37668
Refbacks
- 当前没有refback。

此作品已接受知识共享署名-非商业性使用 4.0国际许可协议的许可。





