面向网络攻击的电力信息物理系统风险量化评估研究综述

针对网络攻击严重威胁电力系统安全运行的现状,文中对面向网络攻击的电力信息物理系统风险量化评估问题展开分析。本文从由网络攻击造成的典型停电事故出发,阐述电力信息物理系统的风险来源及传播机理。然后对比传统信息系统及电力系统,对电力信息物理系统风险量化评估的特点展开讨论,并进一步基于节点风险概率、风险传播概率和物理量损失量三个方面,对电力系统风险量化计算方法进行分析总结。最后,面向网络攻击的发展方向,对未来研究提出建议和展望。In view of the current situation that cyber attacks pose a serious threat to the safe operation of power systems, this paper analyzes the quantitative risk assessment of cyber-physical power systems for cyber attacks. Based on the typical power outages caused by cyber attacks, the sources of risk and the mechanisms of risk propagation within cyber-physical power systems is elucidated. Then, compared with traditional information systems and power systems, the distinctive characteristics of risk quantification assessment for cyber-physical power systems are discussed. Furthermore, based on nodal risk probability, risk propagation probability, and physical quantity loss, the quantitative risk calculation methods for power systems is summarized. Finally, facing the development direction of network attacks, suggestions and prospects for future research are proposed.

微弱目标检测研究综述

微弱目标检测是安全检测中的一项基础性技术,在海面监测、空中微弱目标检测、网络威胁发现等应用场景具有重要实用价值。由于大尺度场景下存在信噪比低、背景变化大等问题,导致微弱目标检测易出现虚警、漏检等问题,微弱目标检测一直是目标检测领域的研究难点和热点。根据检测图像使用的方法不同,微弱目标检测方法分为单帧型和多帧型两大类,单帧型方法的检测速度快,但在复杂场景下表现不佳,而多帧型方法提高了准确性但需要更多资源。未来,在深度学习和人工智能技术的推动下,微弱目标检测不仅将聚焦于提高检测的准确性和效率,还将探索更加复杂的应用场景,如在动态和多变环境中的实时检测,具有广阔的应用前景。